Department of Computer Science and Engineering (AIML)

(R18)
Machine Learning
Lecture Notes

B. Tech III YEAR – I SEM

Prepared by

Mrs.Swapna
( Professor&HOD-CSM)
Dept. CSE(AIML)

Faculty Name : Mrs Swapna Subject Name :ML

 Syllabus

Faculty Name : Mrs Swapna Subject Name :ML

 MACHINE LEARNING

B.Tech. III Year I Sem. L T P C

3 0 0 3
Prerequisites:
1. Data Structures
2. Knowledge on statistical methods

Course Objectives:
 This course explains machine learning techniques such as decision tree learning, Bayesian
learning etc.
 To understand computational learning theory.
 To study the pattern comparison techniques.

Course Outcomes:
 Understand the concepts of computational intelligence like machine learning
 Ability to get the skill to apply machine learning techniques to address the real time problems
in different areas
 Understand the Neural Networks and its usage in machine learning application.

UNIT - I
Introduction - Well-posed learning problems, designing a learning system, Perspectives and issues in
machine learning
Concept learning and the general to specific ordering – introduction, a concept learning task, concept
learning as search, find-S: finding a maximally specific hypothesis, version spaces and the candidate
elimination algorithm, remarks on version spaces and candidate elimination, inductive bias.
Decision Tree Learning – Introduction, decision tree representation, appropriate problems for decision
tree learning, the basic decision tree learning algorithm, hypothesis space search in decision tree
learning, inductive bias in decision tree learning, issues in decision tree learning.

UNIT - II
Artificial Neural Networks-1– Introduction, neural network representation, appropriate problems for
neural network learning, perceptions, multilayer networks and the back-propagation algorithm.
Artificial Neural Networks-2- Remarks on the Back-Propagation algorithm, An illustrative example:
face recognition, advanced topics in artificial neural networks.
Evaluation Hypotheses – Motivation, estimation hypothesis accuracy, basics of sampling theory, a
general approach for deriving confidence intervals, difference in error of two hypotheses, comparing
learning algorithms.

UNIT - III
Bayesian learning – Introduction, Bayes theorem, Bayes theorem and concept learning, Maximum
Likelihood and least squared error hypotheses, maximum likelihood hypotheses for predicting
probabilities, minimum description length principle, Bayes optimal classifier, Gibs algorithm, Naïve
Bayes classifier, an example: learning to classify text, Bayesian belief networks, the EM algorithm.
Computational learning theory – Introduction, probably learning an approximately correct hypothesis,
sample complexity for finite hypothesis space, sample complexity for infinite hypothesis spaces, the
Faculty Name : Mrs Swapna Subject Name :ML
 mistake bound model of learning.
Instance-Based Learning- Introduction, k-nearest neighbour algorithm, locally weighted regression,
radial basis functions, case-based reasoning, remarks on lazy and eager learning.

UNIT- IV
Genetic Algorithms – Motivation, Genetic algorithms, an illustrative example, hypothesis space
search, genetic programming, models of evolution and learning, parallelizing genetic algorithms.

Faculty Name : Mrs Swapna Subject Name :ML

 Learning Sets of Rules – Introduction, sequential covering algorithms, learning rule sets: summary,
learning First-Order rules, learning sets of First-Order rules: FOIL, Induction as inverted deduction,
inverting resolution.
Reinforcement Learning – Introduction, the learning task, Q–learning, non-deterministic, rewards and
actions, temporal difference learning, generalizing from examples, relationship to dynamic
programming.

UNIT - V
Analytical Learning-1- Introduction, learning with perfect domain theories: PROLOG-EBG, remarks
on explanation-based learning, explanation-based learning of search control knowledge.
Analytical Learning-2-Using prior knowledge to alter the search objective, using prior knowledge to
augment search operators.
Combining Inductive and Analytical Learning – Motivation, inductive-analytical approaches to
learning, using prior knowledge to initialize the hypothesis.

TEXT BOOK:
1. Machine Learning – Tom M. Mitchell, - MGH

REFERENCE BOOK:
2. Machine Learning: An Algorithmic Perspective, Stephen Marshland, Taylor & Francis.

Faculty Name : Mrs Swapna Subject Name :ML

 INTRODUCTION

Computer machinery and intelligence:

o 1950: In 1950, Alan Turing published a seminal paper, "Computer Machinery and Intelligence,"
on the topic of artificial intelligence. In his paper, he asked, "Can machines think?"

Machine intelligence in Games:

o 1952: Arthur Samuel, who was the pioneer of machine learning, created a program that helped an
IBM computer to play a checkers game. It performed better more it played.

1959: In 1959, the term "Machine Learning" was first coined by Arthur Samuel.

The term Machine Learning was coined by Arthur Samuel in 1959, an American pioneer in the field of
computer gaming and artificial intelligence, and stated that it gives computers the ability to learn
without being explicitly programmed

.
Ever since computers were invented, we have wondered whether they might be made to learn.

6 UNIT 1 ML Faculty Name:Mrs Swapna

 If we could understand how to program them to learn-to improve automatically with
experience-the impact would be dramatic.
 Imagine computers learning from medical records which treatments are most effective
for new diseases
 Houses learning from experience to optimize energy costs based on the particular usage
patterns of their occupants.
 Personal software assistants learning the evolving interests of their users in order to
highlight especially relevant stories from the online morning newspaper

A successful understanding of how to make computers learn would open up many new uses
of computers and new levels of competence and customization

Some successful applications of machine learning

 Learning to recognize spoken words
 Learning to drive an autonomous vehicle
 Learning to classify new astronomical structures
 Learning to play world-class backgammon

Why is Machine Learning Important?

 Some tasks cannot be defined well, except by examples (e.g., recognizing people).
 Relationships and correlations can be hidden within large amounts of data. Machine
Learning/Data Mining may be able to find these relationships.
 Human designers often produce machines that do not work as well as desired in the
environments in which they are used.
 The amount of knowledge available about certain tasks might be too large for explicit
encoding by humans (e.g., medical diagnostic).
 Environments change over time.
 New knowledge about tasks is constantly being discovered by humans. It may be
difficult to continuously re-design systems “by hand”.
Classification of Machine Learning

machine learning can be classified into three types:

1. Supervised learning
2. Unsupervised learning
3. Reinforcement learning

1. Supervised Learning

Supervised learning is a type of machine learning method in which we provide sample labeled data to
the machine learning system in order to train it, and on that basis, it predicts the output.

7 UNIT 1 ML Faculty Name:Mrs Swapna

The system creates a model using labeled data to understand the datasets and learn about each data, once
the training and processing are done then we test the model by providing a sample data to check whether
it is predicting the exact output or not.

The goal of supervised learning is to map input data with the output data. The supervised learning is
based on supervision, and it is the same as when a student learns things in the supervision of the teacher.
The example of supervised learning

Supervised learning, as the name indicates, has the presence of a supervisor as a teacher. Basically
supervised learning is when we teach or train the machine using data that is well labeled. Which means
some data is already tagged with the correct answer. After that, the machine is provided with a new set of
examples(data) so that the supervised learning algorithm analyses the training data(set of training
examples) and produces a correct outcome from labeled data.

For instance, suppose you are given a basket filled with different kinds of fruits. Now the first step is to
train the machine with all different fruits one by one like this:
 If the shape of the object is rounded and has a depression at the top, is red in color, then it will be
labeled as –Apple.
 If the shape of the object is a long curving cylinder having Green-Yellow color, then it will be labeled as
–Banana.

Now suppose after training the data, you have given a new separate fruit, say Banana from the basket,
and asked to identify it.
Since the machine has already learned the things from previous data and this time has to use it wisely. It
will first classify the fruit with its shape and color and would confirm the fruit name as BANANA and put it
in the Banana category. Thus the machine learns the things from training data(basket containing fruits)
and then applies the knowledge to test data(new fruit).

Supervised learning is classified into two categories of algorithms:

 Classification: A classification problem is when the output variable is a category, such as “Red” or
“blue” or “disease” and “no disease”.
 Regression: A regression problem is when the output variable is a real value, such as “dollars” or
“weight”.
Supervised learning deals with or learns with “labeled” data. This implies that some data is already tagged
with the correct answer.

2) Unsupervised Learning

Unsupervised learning is a learning method in which a machine learns without any supervision.

8 UNIT 1 ML Faculty Name:Mrs Swapna

The training is provided to the machine with the set of data that has not been labeled, classified, or
categorized, and the algorithm needs to act on that data without any supervision. In unsupervised
learning, we don't have a predetermined result.

when an algorithm learns from plain examples without any associated response, leaving to the
algorithm to determine the data patterns on its own. This type of algorithm tends to restructure the data
into something else, such as new features that may represent a class or a new series of un-correlated
values.

It can be further classifieds into two categories of algorithms

o Clustering
o Association

Structure the input data into new features or a group of objects with similar patterns

Reinforcement learning: When you present the algorithm with examples that lack labels, as in
unsupervised learning. However, you can accompany an example with positive or negative feedback
according to the solution the algorithm proposes comes under the category of Reinforcement learning,
which is connected to applications for which the algorithm must make decisions (so the product is
prescriptive, not just descriptive, as in unsupervised learning), and the decisions bear consequences. In
the human world, it is just like learning by trial and error.
Errors help you learn because they have a penalty added (cost, loss of time, regret, pain, and so on),
teaching you that a certain course of action is less likely to succeed than others. An interesting
example of reinforcement learning occurs when computers learn to play video games by themselves.

Reinforcement learning is a feedback-based learning method, in which a learning agent gets a reward for
each right action and gets a penalty for each wrong action. The agent learns automatically with these
feedbacks and improves its performance

Terminologies of Machine Learning

 Model
A model is a specific representation learned from data by applying some machine learning
algorithm. A model is also called hypothesis.

9 UNIT 1 ML Faculty Name:Mrs Swapna

 Feature
A feature is an individual measurable property of our data. A set of numeric features can be
conveniently described by a feature vector. Feature vectors are fed as input to the model. For
example, in order to predict a fruit, there may be features like color, smell, taste, etc.
Note: Choosing informative, discriminating and independent features is a crucial step for effective
algorithms. We generally employ a feature extractor to extract the relevant features from the raw
data.
 Target (Label)
A target variable or label is the value to be predicted by our model. For the fruit example discussed
in the features section, the label with each set of input would be the name of the fruit like apple,
orange, banana, etc.
 Training
The idea is to give a set of inputs(features) and it’s expected outputs(labels), so after training, we
will have a model (hypothesis) that will then map new data to one of the categories trained on.
 Prediction
Once our model is ready, it can be fed a set of inputs to which it will provide a predicted
output(label).

Basic Difference in ML and Traditional Programming?

 Traditional Programming : We feed in DATA (Input) + PROGRAM (logic), run it on machine
and get output.
 Machine Learning : We feed in DATA(Input) + Output, run it on machine during training and the
machine creates its own program(logic), which can be evaluated while testing.

10 UNIT 1 ML Faculty Name:Mrs Swapna

In 1997, Tom Mitchell gave a “well-posed” mathematical and relational definition that “A computer
program is said to learn from experience E with respect to some task T and some performance measure
P, if its performance on T, as measured by P, improves with experience E.

WELL-POSED LEARNING PROBLEMS

Definition: A computer program is said to learn from experience E with respect to some class
of tasks T and performance measure P, if its performance at tasks in T, as measured by P,
improves with experience E.
Example: playing checkers.
E = the experience of playing many games of checkers
T = the task of playing checkers.
P = the probability that the program will win the next game

To have a well-defined learning problem, three features needs to be identified:

1. The class of tasks
2. The measure of performance to be improved
3. The source of experience

Examples
1. Checkers game: A computer program that learns to play checkers might improve its
performance as measured by its ability to win at the class of tasks involving playing
checkers games, through experience obtained by playing games against itself.

Fig: Checker game board

A checkers learning problem:
 Task T: playing checkers
 Performance measure P: percent of games won against opponents
 Training experience E: playing practice games against itself

2. A handwriting recognition learning problem:

 Task T: recognizing and classifying handwritten words within images
 Performance measure P: percent of words correctly classified
 Training experience E: a database of handwritten words with given
11 UNIT 1 ML Faculty Name:Mrs Swapna
 classifications
3. A robot driving learning problem:
 Task T: driving on public four-lane highways using vision sensors
 Performance measure P: average distance travelled before an error.
 Training experience E: a sequence of images and steering commands recorded
while observing a human driver

12 UNIT 1 ML Faculty Name:Mrs Swapna

DESIGNING A LEARNING SYSTEM
The basic design issues and approaches to machine learning are illustrated by designing a
program to learn to play checkers, with the goal of entering it in the world checkers
tournament
1. Choosing the Training Experience
2. Choosing the Target Function
3. Choosing a Representation for the Target Function
4. Choosing a Function Approximation Algorithm
1. Estimating training values
2. Adjusting the weights
5. The Final Design

1. Choosing the Training Experience

 The first design choice is to choose the type of training experience from which the
system will learn.
 The type of training experience available can have a significant impact on success or
failure of the learner.

There are three attributes which impact on success or failure of the learner

1. Whether the training experience provides direct or indirect feedback regarding the
choices made by the performance system.

For example, in checkers game:

In learning to play checkers, the system might learn from direct training examples
consisting of individual checkers board states and the correct move for each.

Indirect training examples consisting of the move sequences and final outcomes of
various games played. The information about the correctness of specific moves early in
the game must be inferred indirectly from the fact that the game was eventually won or
lost.

Here the learner faces an additional problem of credit assignment, or determining the
degree to which each move in the sequence deserves credit or blame for the final
outcome. Credit assignment can be a particularly difficult problem because the game
can be lost even when early moves are optimal, if these are followed later by poor
moves.
Hence, learning from direct training feedback is typically easier than learning from
indirect feedback.

13 UNIT 1 ML Faculty Name:Mrs Swapna

 2. The degree to which the learner controls the sequence of training examples

For example, in checkers game:

The learner might depends on the teacher to select informative board states and to
provide the correct move for each.

Alternatively, the learner might itself propose board states that it finds particularly
confusing and ask the teacher for the correct move.

The learner may have complete control over both the board states and (indirect) training
classifications, as it does when it learns by playing against itself with no teacher present.

3. How well it represents the distribution of examples over which the final system
performance P must be measured

For example, in checkers game:

In checkers learning scenario, the performance metric P is the percent of games the
system wins in the world tournament.

If its training experience E consists only of games played against itself, there is a danger
that this training experience might not be fully representative of the distribution of
situations over which it will later be tested.
It is necessary to learn from a distribution of examples that is different from those on
which the final system will be evaluated.
2. Choosing the Target Function

The next design choice is to determine exactly what type of knowledge will be learned and
how this will be used by the performance program.

Let’s consider a checkers-playing program that can generate the legal moves from any board
state.
The program needs only to learn how to choose the best move from among these legal moves.
We must learn to choose among the legal moves, the most obvious choice for the type of
information to be learned is a program, or function, that chooses the best move for any given
board state.

1. Let ChooseMove be the target function and the notation is

ChooseMove : B→ M
which indicate that this function accepts as input any board from the set of legal board
states B and produces as output some move from the set of legal moves M.
14 UNIT 1 ML Faculty Name:Mrs Swapna
 ChooseMove is a choice for the target function in checkers example, but this function
will turn out to be very difficult to learn given the kind of indirect training experience
available to our system

2. An alternative target function is an evaluation function that assigns a numerical score

to any given board state
Let the target function V and the notation
V:B →R

which denote that V maps any legal board state from the set B to some real value.
Intend for this target function V to assign higher scores to better board states. If the
system can successfully learn such a target function V, then it can easily use it to select
the best move from any current board position.

Let us define the target value V(b) for an arbitrary board state b in B, as follows:
 If b is a final board state that is won, then V(b) = 100
 If b is a final board state that is lost, then V(b) = -100
 If b is a final board state that is drawn, then V(b) = 0

3. Choosing a Representation for the Target Function

Let’s choose a simple representation - for any given board state, the function c will be
calculated as a linear combination of the following board features:

 xl: the number of black pieces on the board

 x2: the number of red pieces on the board
 x3: the number of black kings on the board
 x4: the number of red kings on the board
 x5: the number of black pieces threatened by red (i.e., which can be captured on red's
next turn)
 x6: the number of red pieces threatened by black

Thus, learning program will represent as a linear function of the form

15 UNIT 1 ML Faculty Name:Mrs Swapna

Where,
 w0 through w6 are numerical coefficients, or weights, to be chosen by the learning
algorithm.
 Learned values for the weights w1 through w6 will determine the relative importance
of the various board features in determining the value of the board
 The weight w0 will provide an additive constant to the board value

4. Choosing a Function Approximation Algorithm

In order to learn the target function f we require a set of training examples, each describing a
specific board state b and the training value Vtrain(b) for b.

Each training example is an ordered pair of the form (b, Vtrain(b)).

For instance, the following training example describes a board state b in which black has won
the game (note x2 = 0 indicates that red has no remaining pieces) and for which the target
function value Vtrain(b) is therefore +100.

((x1=3, x2=0, x3=1, x4=0, x5=0, x6=0), +100)

Function Approximation Procedure

1. Derive training examples from the indirect training experience available to the learner
2. Adjusts the weights wi to best fit these training examples

1. Estimating training values

A simple approach for estimating training values for intermediate board states is to
assign the training value of Vtrain(b) for any intermediate board state b to be
(Successor(b))

Where ,
 is the learner's current approximation to V
 Successor(b) denotes the next board state following b for which it is again the
program's turn to move

Rule for estimating training values

Vtrain (Successor(b))

16 UNIT 1 ML Faculty Name:Mrs Swapna

2. Adjusting the weights
Specify the learning algorithm for choosing the weights wi to best fit the set of training
examples {(b, Vtrain(b))}
A first step is to define what we mean by the bestfit to the training data.
One common approach is to define the best hypothesis, or set of weights, as that which
minimizes the squared error E between the training values and the values predicted by
the hypothesis.

Several algorithms are known for finding weights of a linear function that minimize E.
One such algorithm is called the least mean squares, or LMS training rule. For each
observed training example it adjusts the weights a small amount in the direction that
reduces the error on this training example

LMS weight update rule :- For each training example (b, Vtrain(b))
se the current weights to calculate (b)
For each weight wi, update it as

wi ← wi + ƞ ( train (b) - (b)) xi

Here ƞ is a small constant (e.g., 0.1) that moderates the size of the weight update.

Working of weight update rule

 When the error (Vtrain(b)- (b)) is zero, no weights are changed.

 When (Vtrain(b) - (b)) is positive (i.e., when (b) is too low), then each weightis
increased in proportion to the value of its corresponding feature. This will raise
the value of (b), reducing the error.
 If the value of some feature xi is zero, then its weight is not altered regardless of
the error, so that the only weights updated are those whose features actually occur
on the training example board.

17 UNIT 1 ML Faculty Name:Mrs Swapna

5. The Final Design
The final design of checkers learning system can be described by four distinct program modules
that represent the central components in many learning systems

1. The Performance System is the module that must solve the given performance task by
using the learned target function(s). It takes an instance of a new problem (new game)
as input and produces a trace of its solution (game history) as output.

2. The Critic takes as input the history or trace of the game and produces as output a set
of training examples of the target function

3. The Generalizer takes as input the training examples and produces an output
hypothesis that is its estimate of the target function. It generalizes from the specific
training examples, hypothesizing a general function that covers these examples and
other cases beyond the training examples.

4. The Experiment Generator takes as input the current hypothesis and outputs a new
problem (i.e., initial board state) for the Performance System to explore. Its role is to
pick new practice problems that will maximize the learning rate of the overall system.

18 UNIT 1 ML Faculty Name:Mrs Swapna

 PERSPECTIVES AND ISSUES IN MACHINE LEARNING

Issues in Machine Learning

The field of machine learning, and much of this book, is concerned with answering questions
such as the following
 What algorithms exist for learning general target functions from specific training
examples? In what settings will particular algorithms converge to the desired function,
given sufficient training data? Which algorithms perform best for which types of
problems and representations?
 How much training data is sufficient? What general bounds can be found to relate the
confidence in learned hypotheses to the amount of training experience and the character
of the learner's hypothesis space?
 When and how can prior knowledge held by the learner guide the process of generalizing
from examples? Can prior knowledge be helpful even when it is only approximately
correct?
 What is the best strategy for choosing a useful next training experience, and how does
the choice of this strategy alter the complexity of the learning problem?
 What is the best way to reduce the learning task to one or more function approximation
problems? Put another way, what specific functions should the system attempt to learn?
Can this process itself be automated?
 How can the learner automatically alter its representation to improve its ability to
represent and learn the target function?

Issues in Machine Learning

Although machine learning is being used in every industry and helps organizations make more informed
and data-driven choices that are more effective than classical methodologies, it still has so many
problems that cannot be ignored. Here are some common issues in Machine Learning that professionals
face to inculcate ML skills and create an application from scratch.

1. Inadequate Training Data

The major issue that comes while using machine learning algorithms is the lack of quality as well as
quantity of data. Although data plays a vital role in the processing of machine learning algorithms, many
data scientists claim that inadequate data, noisy data, and unclean data are extremely exhausting the
machine learning algorithms. For example, a simple task requires thousands of sample data, and an
advanced task such as speech or image recognition needs millions of sample data examples. Further,

19 UNIT 1 ML Faculty Name:Mrs Swapna

data quality is also important for the algorithms to work ideally, but the absence of data quality is also
found in Machine Learning applications. Data quality can be affected by some factors as follows:

o Noisy Data- It is responsible for an inaccurate prediction that affects the decision as well as
accuracy in classification tasks.
o Incorrect data- It is also responsible for faulty programming and results obtained in machine
learning models. Hence, incorrect data may affect the accuracy of the results also.
o Generalizing of output data- Sometimes, it is also found that generalizing output data becomes
complex, which results in comparatively poor future actions.

2. Poor quality of data

As we have discussed above, data plays a significant role in machine learning, and it must be of good
quality as well. Noisy data, incomplete data, inaccurate data, and unclean data lead to less accuracy in
classification and low-quality results. Hence, data quality can also be considered as a major common
problem while processing machine learning algorithms.

3. Non-representative training data

To make sure our training model is generalized well or not, we have to ensure that sample training data
must be representative of new cases that we need to generalize. The training data must cover all cases
that are already occurred as well as occurring.

Further, if we are using non-representative training data in the model, it results in less accurate
predictions. A machine learning model is said to be ideal if it predicts well for generalized cases and
provides accurate decisions. If there is less training data, then there will be a sampling noise in the
model, called the non-representative training set. It won't be accurate in predictions. To overcome this, it
will be biased against one class or a group.

Hence, we should use representative data in training to protect against being biased and make accurate
predictions without any drift.

4. Overfitting and Underfitting

Overfitting is one of the most common issues faced by Machine Learning engineers and data scientists.
Whenever a machine learning model is trained with a huge amount of data, it starts capturing noise and
inaccurate data into the training data set. It negatively affects the performance of the model. Let's
20 UNIT 1 ML Faculty Name:Mrs Swapna
understand with a simple example where we have a few training data sets such as 1000 mangoes, 1000
apples, 1000 bananas, and 5000 papayas. Then there is a considerable probability of identification of an
apple as papaya because we have a massive amount of biased data in the training data set; hence
prediction got negatively affected.

Underfitting:

Under fitting is just the opposite of over fitting. Whenever a machine learning model is trained with fewer amounts

of data, and as a result, it provides incomplete and inaccurate data and destroys the accuracy of the machine

learning model. Under fitting occurs when our model is too simple to understand the base structure of the data, is
less in quantity.

5. Monitoring and maintenance

As we know that generalized output data is mandatory for any machine learning model; hence, regular
monitoring and maintenance become compulsory for the same. Different results for different actions
require data change; hence editing of codes as well as resources for monitoring them also become
necessary.

6. Getting bad recommendations

A machine learning model operates under a specific context which results in bad recommendations and
concept drift in the model. Let's understand with an example where at a specific time customer is looking
for some gadgets, but now customer requirement changed over time but still machine learning model
showing same recommendations to the customer while customer expectation has been changed. This
incident is called a Data Drift. It generally occurs when new data is introduced or interpretation of data
changes. However, we can overcome this by regularly updating and monitoring data according to the
expectations.

7. Lack of skilled resources

Although Machine Learning and Artificial Intelligence are continuously growing in the market, still
these industries are fresher in comparison to others. The absence of skilled resources in the form of
manpower is also an issue. Hence, we need manpower having in-depth knowledge of mathematics,
science, and technologies for developing and managing scientific substances for machine learning.

8. Customer Segmentation

21 UNIT 1 ML Faculty Name:Mrs Swapna

Customer segmentation is also an important issue while developing a machine learning algorithm. To
identify the customers who paid for the recommendations shown by the model and who don't even check
them. Hence, an algorithm is necessary to recognize the customer behavior and trigger a relevant
recommendation for the user based on past experience.

9. Process Complexity of Machine Learning

The machine learning process is very complex, which is also another major issue faced by machine
learning engineers and data scientists. However, Machine Learning and Artificial Intelligence are very
new technologies but are still in an experimental phase and continuously being changing over time.
There is the majority of hits and trial experiments; hence the probability of error is higher than expected.
Further, it also includes analyzing the data, removing data bias, training data, applying complex
mathematical calculations, etc., making the procedure more complicated and quite tedious.

10. Data Bias

Data Biasing is also found a big challenge in Machine Learning. These errors exist when certain
elements of the dataset are heavily weighted or need more importance than others. Biased data leads to
inaccurate results, skewed outcomes, and other analytical errors. However, we can resolve this error by
determining where data is actually biased in the dataset. Further, take necessary steps to reduce it.

11. Lack of Explainability

This basically means the outputs cannot be easily comprehended as it is programmed in specific ways to
deliver for certain conditions. Hence, a lack of explainability is also found in machine learning
algorithms which reduce the credibility of the algorithms.

12. Slow implementations and results

This issue is also very commonly seen in machine learning models. However, machine learning models
are highly efficient in producing accurate results but are time-consuming. Slow programming, excessive
requirements' and overloaded data take more time to provide accurate results than expected. This needs
continuous maintenance and monitoring of the model for delivering accurate results.

13. Irrelevant features

Although machine learning models are intended to give the best possible outcome, if we feed garbage
data as input, then the result will also be garbage. Hence, we should use relevant features in our training
22 UNIT 1 ML Faculty Name:Mrs Swapna
sample. A machine learning model is said to be good if training data has a good set of features or less to
no irrelevant features.

CONCEPT LEARNING

 Learning involves acquiring general concepts from specific training examples. Example:
People continually learn general concepts or categories such as "bird," "car," "situations in
which I should study more in order to pass the exam," etc.
 Each such concept can be viewed as describing some subset of objects or events defined
over a larger set
 Alternatively, each concept can be thought of as a Boolean-valued function defined over this
larger set. (Example: A function defined over all animals, whose value is true for birds and
false for other animals).

Definition: Concept learning - Inferring a Boolean-valued function from training examples of

its input and output

A CONCEPT LEARNING TASK

Consider the example task of learning the target concept "Days on which Aldo enjoys
his favorite water sport”

Example Sky AirTemp Humidity Wind Water Forecast EnjoySport

1 Sunny Warm Normal Strong Warm Same Yes

2 Sunny Warm High Strong Warm Same Yes

3 Rainy Cold High Strong Warm Change No

4 Sunny Warm High Strong Cool Change Yes

Table: Positive and negative training examples for the target concept EnjoySport.

The task is to learn to predict the value of EnjoySport for an arbitrary day, based on the
values of its other attributes?

What hypothesis representation is provided to the learner?

 Let’s consider a simple representation in which each hypothesis consists of a

conjunction of constraints on the instance attributes.
 Let each hypothesis be a vector of six constraints, specifying the values of the six
23 UNIT 1 ML Faculty Name:Mrs Swapna
attributes Sky, AirTemp, Humidity, Wind, Water, and Forecast.

24 UNIT 1 ML Faculty Name:Mrs Swapna

For each attribute, the hypothesis will either
 Indicate by a "?' that any value is acceptable for this attribute,
 Specify a single required value (e.g., Warm) for the attribute, or
 Indicate by a "Φ" that no value is acceptable

If some instance x satisfies all the constraints of hypothesis h, then h classifies x as a positive
example (h(x) = 1).

The hypothesis that PERSON enjoys his favorite sport only on cold days with high humidity
is represented by the expression
(?, Cold, High, ?, ?, ?)

The most general hypothesis-that every day is a positive example-is represented by

(?, ?, ?, ?, ?, ?)

The most specific possible hypothesis-that no day is a positive example-is represented by

(Φ, Φ, Φ, Φ, Φ, Φ)

Notation

 The set of items over which the concept is defined is called the set of instances, which is
denoted by X.

Example: X is the set of all possible days, each represented by the attributes: Sky, AirTemp,
Humidity, Wind, Water, and Forecast

 The concept or function to be learned is called the target concept, which is denoted by c.
c can be any Boolean valued function defined over the instances X

c: X→ {O, 1}

Example: The target concept corresponds to the value of the attribute EnjoySport
(i.e., c(x) = 1 if EnjoySport = Yes, and c(x) = 0 if EnjoySport = No).

 Instances for which c(x) = 1 are called positive examples, or members of the target concept.
 Instances for which c(x) = 0 are called negative examples, or non-members of the target
concept.
 The ordered pair (x, c(x)) to describe the training example consisting of the instance x and
its target concept value c(x).

25 UNIT 1 ML Faculty Name:Mrs Swapna

 D to denote the set of available training examples

26 UNIT 1 ML Faculty Name:Mrs Swapna

 The symbol H to denote the set of all possible hypotheses that the learner may consider
regarding the identity of the target concept. Each hypothesis h in H represents a Boolean-
valued function defined over X
h: X→{O, 1}

The goal of the learner is to find a hypothesis h such that h(x) = c(x) for all x in X.

 Given:
 Instances X: Possible days, each described by the attributes
 Sky (with possible values Sunny, Cloudy, and Rainy),
 AirTemp (with values Warm and Cold),
 Humidity (with values Normal and High),
 Wind (with values Strong and Weak),
 Water (with values Warm and Cool),
 Forecast (with values Same and Change).

 Hypotheses H: Each hypothesis is described by a conjunction of constraints on the

attributes Sky, AirTemp, Humidity, Wind, Water, and Forecast. The constraints may be
"?" (any value is acceptable), “Φ” (no value is acceptable), or a specific value.

 Target concept c: EnjoySport : X → {0, l}

 Training examples D: Positive and negative examples of the target function

 Determine:
 A hypothesis h in H such that h(x) = c(x) for all x in X.

Table: The EnjoySport concept learning task.

The inductive learning hypothesis

Any hypothesis found to approximate the target function well over a sufficiently large set of
training examples will also approximate the target function well over other unobserved
examples.

27 UNIT 1 ML Faculty Name:Mrs Swapna

CONCEPT LEARNING AS SEARCH

 Concept learning can be viewed as the task of searching through a large space of
hypotheses implicitly defined by the hypothesis representation.
 The goal of this search is to find the hypothesis that best fits the training examples.

Example:
Consider the instances X and hypotheses H in the EnjoySport learning task. The attribute Sky
has three possible values, and AirTemp, Humidity, Wind, Water, Forecast each have two
possible values, the instance space X contains exactly
3.2.2.2.2.2 = 96 distinct instances v

5.4.4.4.4.4 = 5120 distinct hypotheses within H.(specific and general )

1 + (4.3.3.3.3.3) = 973. Semantically distinct hypotheses(only general +1 specific)

FIND-S: FINDING A MAXIMALLY SPECIFIC HYPOTHESIS

FIND-S Algorithm

1. Initialize h to the most specific hypothesis in H

2. For each positive training instance x
For each attribute constraint ai in h
If the constraint ai is satisfied by x
Then do nothing
Else replace ai in h by the next more general constraint that is satisfied by x
3. Output hypothesis h

To illustrate this algorithm, assume the learner is given the sequence of training examples
from the EnjoySport task

Example Sky AirTemp Humidity Wind Water Forecast EnjoySport

1 Sunny Warm Normal Strong Warm Same Yes
2 Sunny Warm High Strong Warm Same Yes
3 Rainy Cold High Strong Warm Change No
4 Sunny Warm High Strong Cool Change Yes

 The first step of FIND-S is to initialize h to the most specific hypothesis in H

28 UNIT 1 ML Faculty Name:Mrs Swapna

 h - (Ø, Ø, Ø, Ø, Ø, Ø)

 Consider the first training example

x1 = <Sunny Warm Normal Strong Warm Same>, +

Observing the first training example, it is clear that hypothesis h is too specific. None
of the "Ø" constraints in h are satisfied by this example, so each is replaced by the next
more general constraint that fits the example
h1 = <Sunny Warm Normal Strong Warm Same>

 Consider the second training example

x2 = <Sunny, Warm, High, Strong, Warm, Same>, +

The second training example forces the algorithm to further generalize h, this time
substituting a "?" in place of any attribute value in h that is not satisfied by the new
example
h2 = <Sunny Warm ? Strong Warm Same>

 Consider the third training example

x3 = <Rainy, Cold, High, Strong, Warm, Change>, -

Upon encountering the third training the algorithm makes no change to h. The FIND-S
algorithm simply ignores every negative example.
h3 = < Sunny Warm ? Strong Warm Same>

 Consider the fourth training example

x4 = <Sunny Warm High Strong Cool Change>, +

The fourth example leads to a further generalization of h

h4 = < Sunny Warm ? Strong ? ? >

29 UNIT 1 ML Faculty Name:Mrs Swapna

 The key property of the FIND-S algorithm
 FIND-S is guaranteed to output the most specific hypothesis within H that is consistent
with the positive training examples
 FIND-S algorithm’s final hypothesis will also be consistent with the negative examples
provided the correct target concept is contained in H, and provided the training examples
are correct.

Unanswered by FIND-S

1. Has the learner converged to the correct target concept?

2. Why prefer the most specific hypothesis?
3. Are the training examples consistent?
4. What if there are several maximally specific consistent hypotheses?

Limitations of Find-S Algorithm

There are a few limitations of the Find-S algorithm listed down below:

1. There is no way to determine if the hypothesis is consistent throughout the data.

2. Inconsistent training sets can actually mislead the Find-S algorithm, since it ignores the negative
examples.

30 UNIT 1 ML Faculty Name:Mrs Swapna

VERSION SPACES AND THE CANDIDATE-ELIMINATION ALGORITHM

The key idea in the CANDIDATE-ELIMINATION algorithm is to output a description of the

set of all hypotheses consistent with the training examples

Representation

Definition: consistent- A hypothesis h is consistent with a set of training examples D if and

only if h(x) = c(x) for each example (x, c(x)) in D.

Consistent (h, D)  ( x, c(x)  D) h(x) = c(x))

Note difference between definitions of consistent and satisfies

 An example x is said to satisfy hypothesis h when h(x) = 1, regardless of whether x is
a positive or negative example of the target concept.
 An example x is said to consistent with hypothesis h iff h(x) = c(x)

Definition: version space- The version space, denoted V S with respect to hypothesis space
H, D
H and training examples D, is the subset of hypotheses from H consistent with the training
examples in D
V S {h  H | Consistent (h, D)}
H, D

A More Compact Representation for Version Spaces

The version space is represented by its most general and least general members. These
members form general and specific boundary sets that delimit the version space within the
partially ordered hypothesis space.

Definition: The general boundary G, with respect to hypothesis space H and training data D,
is the set of maximally general members of H consistent with D

G {g  H | Consistent (g, D)(g'  H)[(g'  g)  Consistent(g', D)]}

g

Definition: The specific boundary S, with respect to hypothesis space H and training data D,
is the set of minimally general (i.e., maximally specific) members of H consistent with D.

S {s  H | Consistent (s, D)(s'  H)[(s  s')  Consistent(s', D)]}

g

31 UNIT 1 ML Faculty Name:Mrs Swapna

32 UNIT 1 ML Faculty Name:Mrs Swapna
CANDIDATE-ELIMINATION Learning Algorithm

The CANDIDATE-ELIMINTION algorithm computes the version space containing all

hypotheses from H that are consistent with an observed sequence of training examples.

Initialize G to the set of maximally general hypotheses in H

Initialize S to the set of maximally specific hypotheses in H
For each training example d, do
• If d is a positive example
• Moves from Specific to general
• Remove from G any hypothesis inconsistent with d
• For each hypothesis s in S that is not consistent with d
• Remove s from S
• Add to S all minimal generalizations h of s such that
• h is consistent with d, and some member of G is more general than h
• Remove from S any hypothesis that is more general than another hypothesis in S

• If d is a negative example
• Moves from general to specific
• Remove from S any hypothesis inconsistent with d
• For each hypothesis g in G that is not consistent with d
• Remove g from G
• Add to G all minimal specializations h of g such that
• h is consistent with d, and some member of S is more specific than h
• Remove from G any hypothesis that is less general than another hypothesis in G

CANDIDATE- ELIMINTION algorithm using version spaces

An Illustrative Example

Example Sky AirTemp Humidity Wind Water Forecast EnjoySport

1 Sunny Warm Normal Strong Warm Same Yes
2 Sunny Warm High Strong Warm Same Yes
3 Rainy Cold High Strong Warm Change No
4 Sunny Warm High Strong Cool Change Yes

33 UNIT 1 ML Faculty Name:Mrs Swapna

 CANDIDATE-ELIMINTION algorithm begins by initializing the version space to the set of
all hypotheses in H;

Initializing the G boundary set to contain the most general hypothesis in H

G0 ?, ?, ?, ?, ?, ?

Initializing the S boundary set to contain the most specific (least general) hypothesis
S0 , , , , , 

 When the second training example is observed, it has a similar effect of generalizing S

34 UNIT 1 ML Faculty Name:Mrs Swapna

Given that there are six attributes that could be specified to specialize G 2, why are there only
three new hypotheses in G3?

 Consider the fourth training example.

35 UNIT 1 ML Faculty Name:Mrs Swapna

  This positive example further generalizes the S boundary of the version space. It also
results in removing one member of the G boundary, because this member fails to
cover the new positive example

After processing these four examples, the boundary sets S4 and G4 delimit the version space
of all hypotheses consistent with the set of incrementally observed training examples.

36 UNIT 1 ML Faculty Name:Mrs Swapna

 INDUCTIVE BIAS: Inductive bias refers to the restrictions that are imposed by the
assumptions made in the learning method.

The fundamental questions for inductive inference

1. What if the target concept is not contained in the hypothesis space how the output is
predicted
2. Can we avoid this difficulty by using a hypothesis space that includes every possible
hypothesis by making more generalize
3. How does the size of this hypothesis space influence the ability of the algorithm to
generalize and make predictions
4. How does the size of the hypothesis space influence the number of training examples
that must be observed and how given hypothesis is correct prediction

These fundamental questions are examined in the context of the CANDIDATE-

ELIMINTION algorithm

A Biased Hypothesis Space:

Which are more Specific-
Deals only with Postive consistent data

 Suppose the target concept is not contained in the hypothesis space H, then obvious
solution is to enrich the hypothesis space to include every possible hypothesis.
 Consider the EnjoySport example in which the hypothesis space is restricted to include
only conjunctions of attribute values. Because of this restriction, the hypothesis space is
unable to represent even simple disjunctive target concepts such as
"Sky = Sunny or Sky = Cloudy."
 The following three training examples of disjunctive hypothesis, the algorithm would
find that there are zero hypotheses in the version space

Sunny Warm Normal Strong Cool Change Y

Cloudy Warm Normal Strong Cool Change Y
Rainy Warm Normal Strong Cool Change N

 If Candidate Elimination algorithm is applied, then it end up with empty Version Space.
After first two training example
S= ? Warm Normal Strong Cool Change

 This new hypothesis is overly general and it incorrectly covers the third negative
training example.

37 UNIT 1 ML Faculty Name:Mrs Swapna

An Unbiased Learner :
Which are more Generalized
Have solution for negative inconsistent data

 The solution to the problem of assuring that the target concept is in the hypothesis space H is
to provide a hypothesis space capable of representing every teachable concept that is
representing every possible subset of the instances X.
 The set of all subsets of a set X is called the power set of X

 In the EnjoySport learning task the size of the instance space X of days described by the
six attributes is 96 instances.
 Thus, there are 296 distinct target concepts that could be defined over this instance space
and learner might be called upon to learn.

Example:
Consider the instances X and hypotheses H in the EnjoySport learning task. The attribute Sky
has three possible values, and AirTemp, Humidity, Wind, Water, Forecast each have two
possible values, the instance space X contains exactly
3.2.2.2.2.2 = 96 distinct instances v

5.4.4.4.4.4 = 5120 distinct hypotheses within H.(specific and general )

1 + (4.3.3.3.3.3) = 973. Semantically distinct hypotheses(only general +1 specific)

The below figure explains Inductive and deductive system

 Modelling inductive systems by equivalent deductive systems.
 The input-output behavior of the CANDIDATE-ELIMINATION algorithm using a
hypothesis space H is identical to that of a deductive theorem prover utilizing the
assertion "H contains the target concept." This assertion is therefore called the inductive
bias of the CANDIDATE-ELIMINATION algorithm.
 Characterizing inductive systems by their inductive bias allows modelling them by their
equivalent deductive systems. This provides a way to compare inductive systems
according to their policies for generalizing beyond the observed training data.

38 UNIT 1 ML Faculty Name:Mrs Swapna

 An Example for general Mathematic resdaoning for inductive and deductive system

Inductive Reasoning: Maximilian is a shelter dog. He is happy. All shelter dogs are happy.
Deductive Reasoning: Maximillian is a shelter dog. All shelter dogs are happy.

39 UNIT 1 ML Faculty Name:Mrs Swapna

 DECISION TREE LEARNING
What is a Decision Tree

• ID3 stands for Iterative Dichotomiser 3

• ID3 is a precursor to the C4.5 Algorithm.

• The ID3 algorithm was invented by Ross Quinlan in 1975

• Used to generate a decision tree from a given data set by employing a top-down,
greedy search, to test each attribute at every node of the tree.

• The resulting tree is used to classify future samples.

It is a tool that has applications spanning several different areas. Decision trees can be used for
classification as well as regression problems. The name itself suggests that it uses a flowchart like a tree
structure to show the predictions that result from a series of feature-based splits. It starts with a root node
and ends with a decision made by leaves.

o Decision Tree is a Supervised learning technique that can be used for both classification and Regression
problems, but mostly it is preferred for solving Classification problems. It is a tree-structured classifier,
where internal nodes represent the features of a dataset, branches represent the decision
rules and each leaf node represents the outcome.
40 UNIT 1 ML Faculty Name:Mrs Swapna
 o In a Decision tree, there are two nodes, which are the Decision Node and Leaf Node. Decision nodes are
used to make any decision and have multiple branches, whereas Leaf nodes are the output of those
decisions and do not contain any further branches.
o The decisions or the test are performed on the basis of features of the given dataset.
o It is a graphical representation for getting all the possible solutions to a problem/decision based on
given conditions.
o It is called a decision tree because, similar to a tree, it starts with the root node, which expands on further
branches and constructs a tree-like structure.

o Example: Suppose there is a candidate who has a job offer and wants to decide whether he should accept
the offer or Not. So, to solve this problem, the decision tree starts with the root node (Salary attribute by
ASM). The root node splits further into the next decision node (distance from the office) and one leaf node
based on the corresponding labels. The next decision node further gets split into one decision node (Cab
facility) and one leaf node. Finally, the decision node splits into two leaf nodes (Accepted offers and Declined
offer). Consider the below diagram:

Decision Tree Terminologies

 Root Node: Root node is from where the decision tree starts. It represents the entire dataset, which further gets
divided into two or more homogeneous sets.
 Leaf Node: Leaf nodes are the final output node, and the tree cannot be segregated further after getting a leaf
node.
 Splitting: Splitting is the process of dividing the decision node/root node into sub-nodes according to the given
conditions.
 Branch/Sub Tree: A tree formed by splitting the tree.
41 UNIT 1 ML Faculty Name:Mrs Swapna
 Pruning: Pruning is the process of removing the unwanted branches from the tree.
 Parent/Child node: The root node of the tree is called the parent node, and other nodes are called the child
nodes.

Entropy: Entropy is a metric to measure the impurity in a given attribute. It specifies randomness in data. Entropy
can be calculated as:

Entropy(s)= -P(yes)log2 P(yes)- P(no) log2 P(no)

Where,

o S= Total number of samples

o P(yes)= probability of yes
o P(no)= probability of no

Attribute Selection Measures

While implementing a Decision tree, the main issue arises that how to select the best attribute for the root node and
for sub-nodes. So, to solve such problems there is a technique which is called as Attribute selection measure or
ASM. By this measurement, we can easily select the best attribute for the nodes of the tree through information
gain

Information Gain
INFORMATION GAIN MEASURES THE EXPECTED REDUCTION IN ENTROPY

 Information gain, is the expected reduction in entropy caused by partitioning the

examples according to this attribute.
 The information gain, Gain(S, A) of an attribute A, relative to a collection of examples
S, is defined as

Information Gain:
o Information gain is the measurement of changes in entropy after the segmentation of a dataset based on an
attribute.
o It calculates how much information a feature provides us about a class.
o According to the value of information gain, we split the node and build the decision tree.
o A decision tree algorithm always tries to maximize the value of information gain, and a node/attribute having
the highest information gain is split first. It can be calculated using the below formula:

42 UNIT 1 ML Faculty Name:Mrs Swapna

 o Information Gain= Entropy(S)- [(Weighted Avg) *Entropy(each feature)

Advantages of the Decision Tree

o It is simple to understand as it follows the same process which a human follow while making any decision in
real-life.
o It can be very useful for solving decision-related problems.
o It helps to think about all the possible outcomes for a problem.
o There is less requirement of data cleaning compared to other algorithms.

Disadvantages of the Decision Tree

o The decision tree contains lots of layers, which makes it complex.
o It may have an overfitting issue, which can be resolved using the Random Forest algorithm(not in
syllabus)
o For more class labels, the computational complexity of the decision tree may increase.

ENTROPY MEASURES HOMOGENEITY OF EXAMPLES

To define information gain, we begin by defining a measure called entropy. Entropy

measures the impurity of a collection of examples.

Given a collection S, containing positive and negative examples of some target concept, the
entropy of S relative to this Boolean classification is

Where,
p+ is the proportion of positive examples in S
p- is the proportion of negative examples in S.

43 UNIT 1 ML Faculty Name:Mrs Swapna

Example:
Suppose S is a collection of 14 examples of some boolean concept, including 9 positive and 5
negative examples. Then the entropy of S relative to this boolean classification is

 The entropy is 0 if all members of S belong to the same class

 The entropy is 1 when the collection contains an equal number of positive and negative
examples
 If the collection contains unequal numbers of positive and negative examples, the
entropy is between 0 and 1

DECISION TREE REPRESENTATION

 Decision trees classify instances by sorting them down the tree from the root to some
leaf node, which provides the classification of the instance.
 Each node in the tree specifies a test of some attribute of the instance, and each branch
descending from that node corresponds to one of the possible values for this attribute.
 An instance is classified by starting at the root node of the tree, testing the attribute
specified by this node, then moving down the tree branch corresponding to the value of
the attribute in the given example. This process is then repeated for the subtree rooted
at the new node.

44 UNIT 1 ML Faculty Name:Mrs Swapna

FIGURE: A decision tree for the concept PlayTennis. An example is classified by sorting it
through the tree to the appropriate leaf node, then returning the classification associated with
this leaf

An Illustrative Example

 To illustrate the operation of ID3, consider the learning task represented by the training
examples of below table.
 Here the target attribute PlayTennis, which can have values yes or no for different days.
 Consider the first step through the algorithm, in which the topmost node of the decision
tree is created.

Day Outlook Temperature Humidity Wind PlayTennis

D1 Sunny Hot High Weak No
D2 Sunny Hot High Strong No
D3 Overcast Hot High Weak Yes
D4 Rain Mild High Weak Yes
D5 Rain Cool Normal Weak Yes
D6 Rain Cool Normal Strong No
D7 Overcast Cool Normal Strong Yes
D8 Sunny Mild High Weak No
D9 Sunny Cool Normal Weak Yes
D10 Rain Mild Normal Weak Yes
45 UNIT 1 ML Faculty Name:Mrs Swapna
 D11 Sunny Mild Normal Strong Yes
D12 Overcast Mild High Strong Yes
D13 Overcast Hot Normal Weak Yes
D14 Rain Mild High Strong No

 ID3 determines the information gain for each candidate attribute (i.e., Outlook,
Temperature, Humidity, and Wind), then selects the one with highest information gain.

What are the steps in ID3 algorithm?

The steps in ID3 algorithm are as follows:

1. Calculate entropy for dataset.

2. For each attribute/feature.
2.1. Calculate entropy for all its categorical values.
2.2. Calculate information gain for the feature.
3. Find the feature with maximum information gain.
4. Repeat it until we get the desired tree.

Complete entropy of dataset is:

H(S) = - p(yes) * log2(p(yes)) - p(no) * log2(p(no))

= - (9/14) * log2(9/14) - (5/14) * log2(5/14)
= - (-0.41) - (-0.53)
= 0.94

First Attribute – Outlook

Categorical values - sunny, overcast and rain
H(Outlook=sunny) = -(2/5)*log(2/5)-(3/5)*log(3/5) =0.971
H(Outlook=rain) = -(3/5)*log(3/5)-(2/5)*log(2/5) =0.971
H(Outlook=overcast) = -(4/4)*log(4/4)-0 = 0

Average Entropy Information for Outlook -

I(Outlook) = p(sunny) * H(Outlook=sunny) + p(rain) * H(Outlook=rain) + p(overcast) *
H(Outlook=overcast)
= (5/14)*0.971 + (5/14)*0.971 + (4/14)*0
= 0.693

Information Gain = H(S) - I(Outlook)

= 0.94 - 0.693
= 0.247

Second Attribute - Temperature

Categorical values - hot, mild, cool
H(Temperature=hot) = -(2/4)*log(2/4)-(2/4)*log(2/4) = 1
H(Temperature=cool) = -(3/4)*log(3/4)-(1/4)*log(1/4) = 0.811

46 UNIT 1 ML Faculty Name:Mrs Swapna

H(Temperature=mild) = -(4/6)*log(4/6)-(2/6)*log(2/6) = 0.9179

Average Entropy Information for Temperature -

I(Temperature) = p(hot)*H(Temperature=hot) + p(mild)*H(Temperature=mild) +
p(cool)*H(Temperature=cool)
= (4/14)*1 + (6/14)*0.9179 + (4/14)*0.811
= 0.9108

Information Gain = H(S) - I(Temperature)

= 0.94 - 0.9108
= 0.0292

Third Attribute - Humidity

Categorical values - high, normal
H(Humidity=high) = -(3/7)*log(3/7)-(4/7)*log(4/7) = 0.983
H(Humidity=normal) = -(6/7)*log(6/7)-(1/7)*log(1/7) = 0.591

Average Entropy Information for Humidity -

I(Humidity) = p(high)*H(Humidity=high) + p(normal)*H(Humidity=normal)
= (7/14)*0.983 + (7/14)*0.591
= 0.787

Information Gain = H(S) - I(Humidity)

= 0.94 - 0.787
= 0.153

Fourth Attribute - Wind

Categorical values - weak, strong
H(Wind=weak) = -(6/8)*log(6/8)-(2/8)*log(2/8) = 0.811
H(Wind=strong) = -(3/6)*log(3/6)-(3/6)*log(3/6) = 1

Average Entropy Information for Wind -

I(Wind) = p(weak)*H(Wind=weak) + p(strong)*H(Wind=strong)
= (8/14)*0.811 + (6/14)*1
= 0.892

Information Gain = H(S) - I(Wind)

= 0.94 - 0.892
= 0.048

Here, the attribute with maximum information gain is Outlook. So, the decision tree built so far -

47 UNIT 1 ML Faculty Name:Mrs Swapna

  when Outlook == overcast, it is of pure class(Yes).
Now, we have to repeat same procedure for the data with rows consist of Outlook value as
Sunny and then for Outlook value as Rain. The information gain values for all four
attributes are
Gain(S, Outlook) = 0.246
Gain(S, Humidity) = 0.151
Gain(S, Wind) = 0.048
Gain(S, Temperature) = 0.029

 According to the information gain measure, the Outlook attribute provides the best
prediction of the target attribute, PlayTennis, over the training examples. Therefore,
Outlook is selected as the decision attribute for the root node, and branches are created
below the root for each of its possible values i.e., Sunny, Overcast, and Rain.

finding the best attribute for splitting the data with Outlook=Sunny

Complete entropy of Sunny is -

48 UNIT 1 ML Faculty Name:Mrs Swapna

H(S) = - p(yes) * log2(p(yes)) - p(no) * log2(p(no))
= - (2/5) * log2(2/5) - (3/5) * log2(3/5)
= 0.971

First Attribute - Temperature

Categorical values - hot, mild, cool

H(Sunny, Temperature=hot) = -0-(2/2)*log(2/2) = 0

H(Sunny, Temperature=cool) = -(1)*log(1)- 0 = 0

H(Sunny, Temperature=mild) = -(1/2)*log(1/2)-(1/2)*log(1/2) = 1
Average Entropy Information for Temperature -
I(Sunny, Temperature) = p(Sunny, hot)*H(Sunny, Temperature=hot) + p(Sunny,
mild)*H(Sunny, Temperature=mild) + p(Sunny, cool)*H(Sunny, Temperature=cool)
= (2/5)*0 + (1/5)*0 + (2/5)*1
= 0.4

Information Gain = H(Sunny) - I(Sunny, Temperature)

= 0.971 - 0.4
= 0.571

Second Attribute - Humidity

Categorical values - high, normal

H(Sunny, Humidity=high) = - 0 - (3/3)*log(3/3) = 0

H(Sunny, Humidity=normal) = -(2/2)*log(2/2)-0 = 0

Average Entropy Information for Humidity -

I(Sunny, Humidity) = p(Sunny, high)*H(Sunny, Humidity=high) + p(Sunny,
normal)*H(Sunny, Humidity=normal)
= (3/5)*0 + (2/5)*0
= 0

Information Gain = H(Sunny) - I(Sunny, Humidity)

= 0.971 - 0
= 0.971

Third Attribute – Wind

49 UNIT 1 ML Faculty Name:Mrs Swapna

Categorical values - weak, strong
H(Sunny, Wind=weak) = -(1/3)*log(1/3)-(2/3)*log(2/3) = 0.918
H(Sunny, Wind=strong) = -(1/2)*log(1/2)-(1/2)*log(1/2) = 1

Average Entropy Information for Wind -

I(Sunny, Wind) = p(Sunny, weak)*H(Sunny, Wind=weak) + p(Sunny, strong)*H(Sunny, Wind=strong)
= (3/5)*0.918 + (2/5)*1
= 0.9508

Information Gain = H(Sunny) - I(Sunny, Wind)

= 0.971 - 0.9508
= 0.0202

Here, the attribute with maximum information gain is Humidity. So, the decision tree built so far -

Here, when Outlook = Sunny and Humidity = High, it is a pure class of category "no". And When
Outlook = Sunny and Humidity = Normal, it is again a pure class of category "yes". Therefore, we
don't need to do further calculations.

Complete entropy of Rain is -

H(S) = - p(yes) * log2(p(yes)) - p(no) * log2(p(no))
= - (3/5) * log(3/5) - (2/5) * log(2/5)
= 0.971

First Attribute - Temperature

50 UNIT 1 ML Faculty Name:Mrs Swapna

 First Attribute - Temperature

Categorical values - mild, cool

H(Rain, Temperature=cool) = -(1/2)*log(1/2)- (1/2)*log(1/2) = 1

H(Rain, Temperature=mild) = -(2/3)*log(2/3)-(1/3)*log(1/3) = 0.918

Average Entropy Information for Temperature -
I(Rain, Temperature) = p(Rain, mild)*H(Rain, Temperature=mild) + p(Rain, cool)*H(Rain,
Temperature=cool)
= (2/5)*1 + (3/5)*0.918
= 0.9508

Information Gain = H(Rain) - I(Rain, Temperature)

= 0.971 - 0.9508
= 0.0202

Second Attribute - Wind

Categorical values - weak, strong

H(Wind=weak) = -(3/3)*log(3/3)-0 = 0
H(Wind=strong) = 0-(2/2)*log(2/2) = 0

Average Entropy Information for Wind -

I(Wind) = p(Rain, weak)*H(Rain, Wind=weak) + p(Rain, strong)*H(Rain, Wind=strong)
= (3/5)*0 + (2/5)*0
= 0

Information Gain = H(Rain) - I(Rain, Wind)

= 0.971 - 0
= 0.971

Here, the attribute with maximum information gain is Wind. So, the decision tree built so far -

51 UNIT 1 ML Faculty Name:Mrs Swapna

Here, when Outlook = Rain and Wind = Strong, it is a pure class of category "no". And When Outlook
= Rain and Wind = Weak, it is again a pure class of category "yes".
And this is our final desired tree for the given dataset.

REAL time implementation of Algorithm

ID3 Algorithm is used to build a Decision Tree to predict the weather.

HYPOTHESIS SPACE SEARCH IN DECISION TREE

LEARNING
 ID3 can be characterized as searching a space of hypotheses for one that fits the training
examples.
 The hypothesis space searched by ID3 is the set of possible decision trees.
 ID3 performs a simple-to complex, hill-climbing search through this hypothesis space,
beginning with the empty tree, then considering progressively more elaborate
hypotheses in search of a decision tree that correctly classifies the training data

52 UNIT 1 ML Faculty Name:Mrs Swapna

 Figure: Hypothesis space search by ID3. ID3 searches through the space of possible decision
trees from simplest to increasingly complex, guided by the information gain.
ID3 in terms of its search space and search strategy, there are some insight into itscapabilities
and limitations

1. ID3's hypothesis space of all decision trees is a complete space of finite discrete-valued
functions, relative to the available attributes. Because every finite discrete-valued
function can be represented by some decision tree
ID3 avoids one of the major risks of methods that search incomplete hypothesis spaces:
that the hypothesis space might not contain the target function.

53 UNIT 1 ML Faculty Name:Mrs Swapna

 2. ID3 maintains only a single current hypothesis as it searches through the space of
decision trees.

For example, with the earlier version space candidate elimination method, which
maintains the set of all hypotheses consistent with the available training examples. But
ID3 maintains Single set of Hypothesis formed

3. ID3 in its pure form performs no backtracking in its search. Once it selects an attribute
to test at a particular level in the tree, it never backtracks to reconsider this choice.
In the case of ID3, corresponds to the decision tree it selects along the single search
path it explores.

4. ID3 uses all training examples at each step in the search to make statistically based
decisions regarding how to refine its current hypothesis.

Why Prefer Short trees

Hypotheses proposed by

Theory of Occam's razor

 Occam's razor: is the problem-solving principle that the simplest solution tends to be
the right one. When presented with competing hypotheses to solve a problem, one
should select the solution with the fewest assumptions.

 Occam's razor: “Prefer the simplest hypothesis that fits the data”.

54 UNIT 1 ML Faculty Name:Mrs Swapna

 INDUCTIVE BIAS IN DECISION TREE LEARNING
Inductive bias is the set of assumptions that, together with the training data, can also
deductively justifythe classifications assigned by the learner to future instances

Given a collection of training examples, there are typically many decision trees consistent with
these examples. Which of these decision trees does ID3 choose?

ID3 search strategy

 Selects in favour of shorter trees over longer ones
 Selects trees that place the attributes with highest information gain closest to the root.

Approximate inductive bias of ID3: Shorter trees are preferred over larger trees

 Consider an algorithm that begins with the empty tree and searches breadth first through
progressively more complex trees.
 First considering all trees of depth 1, then all trees of depth 2, etc.
 Once it finds a decision tree consistent with the training data, it returns the smallest
consistent tree at that search depth (e.g., the tree with the fewest nodes).
 Let us call this breadth-first search algorithm BFS-ID3.
 BFS-ID3 finds a shortest decision tree and thus exhibits the bias "shorter trees are
preferred over longer trees.

A closer approximation to the inductive bias of ID3: Shorter trees are preferred over longer
trees. Trees that place high information gain attributes close to the root are preferred over
those that do not.

 ID3 can be viewed as an efficient approximation to BFS-ID3, using a greedy heuristic

search to attempt to find the shortest tree without conducting the entire breadth-first
search through the hypothesis space.
 Because ID3 uses the information gain heuristic and a hill climbing strategy, it exhibits
a more complex bias than BFS-ID3.

55 UNIT 1 ML Faculty Name:Mrs Swapna

 Restriction Biases and Preference Biases
Difference between the types of inductive bias exhibited by ID3 and by the CANDIDATE-
ELIMINATION Algorithm.
ID3:
 ID3 searches a complete hypothesis space
 It searches incompletely through this space, from simple to complex hypotheses, until
its termination condition is met
 Its inductive bias is solely a consequence of the ordering of hypotheses by its search
strategy.

Preference bias – The inductive bias of ID3 is a preference for certain hypotheses over others
(e.g., preference for shorter hypotheses over larger hypotheses), with no hard restriction on the
hypotheses that can be eventually enumerated. This form of bias is called a preference bias or
a search bias.

Id3 inductive bias is called as Preference bias or search bias

CANDIDATE-ELIMINATION Algorithm:
 It searches this space completely, finding every hypothesis consistent with the training
data.
 Its inductive bias is solely a consequence of the expressive power 2(^96) of its
hypothesis representation.

Restriction bias – The bias of the CANDIDATE ELIMINATION algorithm is in the form of a
categorical restriction on the set of hypotheses considered. This form of bias is typically called
a restriction bias or a language bias.

Candidate Elimaination Algorithm inductive bias is called as Preference bias or search

bias

56 UNIT 1 ML Faculty Name:Mrs Swapna

57 UNIT 1 ML Faculty Name:Mrs Swapna
 Which type of inductive bias is preferred in order to generalize beyond the training data, a
preference bias or restriction bias?

 A preference bias is more desirable than a restriction bias, because it allows the learner
to work within a complete hypothesis space that is assured to contain the unknown target
function.
 In contrast, a restriction bias that strictly limits the set of potential hypotheses is
generally less desirable, because it introduces the possibility of excluding the unknown
target function altogether.

ISSUES IN DECISION TREE LEARNING

Issues in learning decision trees include
1. Avoiding Overfitting the Data
Reduced error pruning
Rule post-pruning
2. Incorporating Continuous-Valued Attributes
3. Alternative Measures for Selecting Attributes
4. Handling Training Examples with Missing Attribute Values
5. Handling Attributes with Differing Costs

1. Avoiding Overfitting the Data

 The ID3 algorithm grows each branch of the tree just deeply enough to perfectly classify the
training examples but it can lead to difficulties when there is noise in the data, or when
the number of training examples is too small to produce a representative sample of the
true target function. This algorithm can produce trees that overfit the training examples.

 Definition - Overfit: Given a hypothesis space H, a hypothesis h ∈ H is said to overfit

the training data if there exists some alternative hypothesis h' ∈ H, such that h has
smaller error than h' over the training examples, but h' has a smaller error than h over
the entire distribution of instances.

58 UNIT 1 ML Faculty Name:Mrs Swapna

The below figure illustrates the impact of overfitting in a typical application of decision tree
learning.

 The horizontal axis of this plot indicates the total number of nodes in the decision tree,
as the tree is being constructed. The vertical axis indicates the accuracy of predictions
made by the tree.
 The solid line shows the accuracy of the decision tree over the training examples. The
broken line shows accuracy measured over an independent set of test example
 The accuracy of the tree over the training examples increases monotonically as the tree
is grown. The accuracy measured over the independent test examples first increases,
then decreases.

How can it be possible for tree h to fit the training examples better than h', but for it to perform
more poorly over subsequent examples?
1. Overfitting can occur when the training examples contain random errors or noise
2. When small numbers of examples are associated with leaf nodes.

Noisy Training Example

 Example 15: <Sunny, Hot, Normal, Strong, ->

 Example is noisy because the correct label is +
 Previously constructed tree misclassifies it

59 UNIT 1 ML Faculty Name:Mrs Swapna

Approaches to avoiding overfitting in decision tree learning
 Pre-pruning (avoidance): Stop growing the tree earlier, before it reaches the point whereit
perfectly classifies the training data
 Post-pruning (recovery): Allow the tree to overfit the data, and then post-prune the tree

Criterion used to determine the correct final tree size

 Use a separate set of examples, distinct from the training examples, to evaluate the utility
of post-pruning nodes from the tree
 Use all the available data for training, but apply a statistical test to estimate whether
expanding (or pruning) a particular node is likely to produce an improvement beyond
the training set
 Use measure of the complexity for encoding the training examples and the decision tree,
halting growth of the tree when this encoding size is minimized. This approach is called
the Minimum Description Length

MDL – Minimize : size(tree) + size (misclassifications(tree))

60 UNIT 1 ML Faculty Name:Mrs Swapna

 Reduced-Error Pruning

 Reduced-error pruning, is to consider each of the decision nodes in the tree to be

candidates for pruning
 Pruning a decision node consists of removing the subtree rooted at that node, making it
a leaf node, and assigning it the most common classification of the training examples
affiliated with that node
 Nodes are removed only if the resulting pruned tree performs no worse than-the original
over the validation set.
 Reduced error pruning has the effect that any leaf node added due to coincidental
regularities in the training set is likely to be pruned because these same coincidences are
unlikely to occur in the validation set

The impact of reduced-error pruning on the accuracy of the decision tree is illustrated in below
figure

 The additional line in figure shows accuracy over the test examples as the tree is pruned.
When pruning begins, the tree is at its maximum size and lowest accuracy over the test
set. As pruning proceeds, the number of nodes is reduced and accuracy over the test set
increases.
 The available data has been split into three subsets: the training examples, the validation
examples used for pruning the tree, and a set of test examples used to provide an
unbiased estimate of accuracy over future unseen examples. The plot shows accuracy
over the training and test sets.

61 UNIT 1 ML Faculty Name:Mrs Swapna

 Rule Post-Pruning
Rule post-pruning is successful method for finding high accuracy hypotheses

 Rule post-pruning involves the following steps:

 Infer the decision tree from the training set, growing the tree until the training data is fit
as well as possible and allowing overfitting to occur.
 Convert the learned tree into an equivalent set of rules by creating one rule for each path
from the root node to a leaf node.
 Prune (generalize) each rule by removing any preconditions that result in improving its
estimated accuracy.
 Sort the pruned rules by their estimated accuracy, and consider them in this sequence
when classifying subsequent instances.

Converting a Decision Tree into Rules

62 UNIT 1 ML Faculty Name:Mrs Swapna

For example, consider the decision tree. The leftmost path of the tree in below figure is
translated into the rule.
IF (Outlook = Sunny) ^ (Humidity = High)
THEN PlayTennis = No

Given the above rule, rule post-pruning would consider removing the preconditions
(Outlook = Sunny) and (Humidity = High)

 It would select whichever of these pruning steps produced the greatest improvement in
estimated rule accuracy, then consider pruning the second precondition as a further
pruning step.
 No pruning step is performed if it reduces the estimated rule accuracy.

2. Incorporating Continuous-Valued Attributes

Continuous-valued decision attributes can be incorporated into the learned tree.

There are two methods for Handling Continuous Attributes

1. Define new discrete valued attributes that partition the continuous attribute value into a
discrete set of intervals.
E.g., {high ≡ Temp > 35º C, med ≡ 10º C < Temp ≤ 35º C, low ≡ Temp ≤ 10º C}

2. Using thresholds for splitting nodes

e.g., A ≤ a produces subsets A ≤ a and A > a

63 UNIT 1 ML Faculty Name:Mrs Swapna

What threshold-based Boolean attribute should be defined based on Temperature?

 Pick a threshold, c, that produces the greatest information gain

 In the current example, there are two candidate thresholds, corresponding to the values
of Temperature at which the value of PlayTennis changes: (48 + 60)/2, and (80 + 90)/2.
 The information gain can then be computed for each of the candidate attributes,
Temperature >54, and Temperature >85 and the best can be selected (Temperature >54)

3. Alternative Measures for Selecting Attributes

 The problem is if attributes with many values, Gain will select it ?

 Example: consider the attribute Date, which has a very large number of possible values.
(e.g., March 4, 1979).
 If this attribute is added to the PlayTennis data, it would have the highest information
gain of any of the attributes. This is because Date alone perfectly predicts the target
attribute over the training data. Thus, it would be selected as the decision attribute for
the root node of the tree and lead to a tree of depth one, which perfectly classifies the
training data.
 This decision tree with root node Date is not a useful predictor because it perfectly
separates the training data, but poorly predict on subsequent examples.

One Approach: Use GainRatio instead of Gain

The gain ratio measure penalizes attributes by incorporating a split information, that is sensitive
to how broadly and uniformly the attribute splits the data

Where, Si is subset of S, for which attribute A has value vi

64 UNIT 1 ML Faculty Name:Mrs Swapna

4. Handling Training Examples with Missing Attribute Values

The data which is available may contain missing values for some attributes
Example: Medical diagnosis
 <Fever = true, Blood-Pressure = normal, …, Blood-Test = ?, …>
 Sometimes values truly unknown, sometimes low priority (or cost too high)

Strategies for dealing with the missing attribute value

 If node n test A, assign most common value of A among other training examples
sortedto node n
 Assign most common value of A among other training examples with same target value
 Assign a probability pi to each of the possible values vi of A rather than simply
assigning the most common value to A(x)

5. Handling Attributes with Differing Costs

 In some learning tasks the instance attributes may have associated costs.
 For example: In learning to classify medical diseases, the patients described in terms
ofattributes such as Temperature, BiopsyResult, Pulse, BloodTestResults, etc.
 These attributes vary significantly in their costs, both in terms of monetary cost and
cost to patient comfort
 Decision trees use low-cost attributes where possible, and depends only on high-
cost attributes only when needed to produce reliable and accurate classifications

65 UNIT 1 ML Faculty Name:Mrs Swapna

 Unit 2

The term "Artificial neural network" refers to a biologically inspired sub-field of artificial intelligence
modeled after the brain. An Artificial neural network is usually a computational network based on
biological neural networks that construct the structure of the human brain. Similar to a human brain has
neurons interconnected to each other, artificial neural networks also have neurons that are linked to each
other in various layers of the networks. These neurons are known as nodes.

What is Artificial Neural Network

The term "Artificial Neural Network" is derived from Biological neural networks that develop the
structure of a human brain. Similar to the human brain that has neurons interconnected to one another,
artificial neural networks also have neurons that are interconnected to one another in various layers of the
networks. These neurons are known as nodes.

The given figure illustrates the typical diagram of Biological Neural Network.

The typical Artificial Neural Network looks something like the given figure.

Dendrites from Biological Neural Network represent inputs in Artificial Neural Networks, cell nucleus
represents Nodes, synapse represents Weights, and Axon represents Output.

Relationship between Biological neural network and artificial neural network:

Faculty Name : Mrs Swapna Subject Name :ML

Biological Neural Network Artificial Neural Network
Dendrites Inputs
Cell nucleus Nodes
Synapse Weights
Axon Output

An Artificial Neural Network in the field of Artificial intelligence where it attempts to mimic the
network of neurons makes up a human brain so that computers will have an option to understand things and
make decisions in a human-like manner. The artificial neural network is designed by programming
computers to behave simply like interconnected brain cells.

There are around 1000 billion neurons in the human brain. Each neuron has an association point somewhere
in the range of 1,000 and 100,000. In the human brain, data is stored in such a manner as to be distributed,
and we can extract more than one piece of this data when necessary from our memory parallelly. We can
say that the human brain is made up of incredibly amazing parallel processors.

We can understand the artificial neural network with an example, consider an example of a digital logic gate
that takes an input and gives an output. "OR" gate, which takes two inputs. If one or both the inputs are
"On," then we get "On" in output. If both the inputs are "Off," then we get "Off" in output. Here the output
depends upon input. Our brain does not perform the same task. The outputs to inputs relationship keep
changing because of the neurons in our brain, which are "learning."

The architecture of an artificial neural network:

To understand the concept of the architecture of an artificial neural network, we have to understand what a
neural network consists of. In order to define a neural network that consists of a large number of artificial
neurons, which are termed units arranged in a sequence of layers. Lets us look at various types of layers
available in an artificial neural network.

Artificial Neural Network primarily consists of three layers:

Faculty Name : Mrs Swapna Subject Name :ML

Input Layer:

As the name suggests, it accepts inputs in several different formats provided by the programmer.

Hidden Layer:

The hidden layer presents in-between input and output layers. It performs all the calculations to find hidden
features and patterns.

Output Layer:

The input goes through a series of transformations using the hidden layer, which finally results in output
that is conveyed using this layer.

The artificial neural network takes input and computes the weighted sum of the inputs and includes a bias.
This computation is represented in the form of a transfer function.

It determines weighted total is passed as an input to an activation function to produce the output. Activation
functions choose whether a node should fire or not. Only those who are fired make it to the output layer.
There are distinctive activation functions available that can be applied upon the sort of task we are
performing.

Advantages of Artificial Neural Network (ANN)

Faculty Name : Mrs Swapna Subject Name :ML

Parallel processing capability:

Artificial neural networks have a numerical value that can perform more than one task simultaneously.

Capability to work with incomplete knowledge:

After ANN training, the information may produce output even with inadequate data. The loss of
performance here relies upon the significance of missing data.

Having fault tolerance:

Extortion of one or more cells of ANN does not prohibit it from generating output, and this feature makes
the network fault-tolerance.

Disadvantages of Artificial Neural Network:

Assurance of proper network structure:

There is no particular guideline for determining the structure of artificial neural networks. The appropriate
network structure is accomplished through experience, trial, and error.

Unrecognized behavior of the network:

It is the most significant issue of ANN. When ANN produces a testing solution, it does not provide insight
concerning why and how. It decreases trust in the network.

Hardware dependence:

Artificial neural networks need processors with parallel processing power, as per their structure. Therefore,
the realization of the equipment is dependent.

Difficulty of showing the issue to the network:

ANNs can work with numerical data. Problems must be converted into numerical values before being
introduced to ANN. The presentation mechanism to be resolved here will directly impact the performance
of the network. It relies on the user's abilities.

The duration of the network is unknown:

The network is reduced to a specific value of the error, and this value does not give us optimum results.

PERCEPTRON

One type of ANN system is based on a unit called a perceptron. Perceptron is a single layer
neural network.

Faculty Name : Mrs Swapna Subject Name :ML

 Figure: A perceptron

A perceptron takes a vector of real-valued inputs, calculates a linear combination of these

inputs, then outputs a 1 if the result is greater than some threshold and -1 otherwise.
Given inputs x through x, the output O(x1, . . . , xn) computed by the perceptron is

Where, each wi is a real-valued constant, or weight, that determines the contribution of input
xi to the perceptron output.
-w0 is a threshold that the weighted combination of inputs w1x1 + . . . + wnxn must surpass in
order for the perceptron to output a 1.

Faculty Name : Mrs Swapna Subject Name :ML

Representational Power of Perceptrons

The perceptron can be viewed as representing a hyperplane decision surface in the n-

dimensional space of instances (i.e., points)
The perceptron outputs a 1 for instances lying on one side of the hyperplane and outputs a -1
for instances lying on the other side, as illustrated in below figure

Perceptrons can represent all of the primitive Boolean functions AND, OR, NAND (~ AND), and
NOR (~OR)
Example: Representation of AND functions

If A=0 & B=0 → 0*0.6 + 0*0.6 = 0.

This is not greater than the threshold of 1, so the output = 0.
If A=0 & B=1 → 0*0.6 + 1*0.6 = 0.6.
This is not greater than the threshold, so the output = 0.
If A=1 & B=0 → 1*0.6 + 0*0.6 = 0.6.
This is not greater than the threshold, so the output = 0.
If A=1 & B=1 → 1*0.6 + 1*0.6 = 1.2.
This exceeds the threshold, so the output = 1.

Faculty Name : Mrs Swapna Subject Name :ML

 Drawback of perceptron
The perceptron rule finds a successful weight vector when the training examples are linearly
separable, it can fail to converge if the examples are not linearly separable

The Perceptron Training Rule

The learning problem is to determine a weight vector that causes the perceptron to produce the correct
+ 1 or - 1 output for each of the given training examples.

To learn an acceptable weight vector

Begin with random weights, then iteratively apply the perceptron to each training example,
modifying the perceptron weights whenever it misclassifies an example.
This process is repeated, iterating through the training examples as many times as needed
until the perceptron classifies all training examples correctly.
Weights are modified at each step according to the perceptron training rule, which revises the
weight wi associated with input xi according to the rule.

Drawback:
The perceptron rule finds a successful weight vector when the training examples are linearly separable, it
can fail to converge if the examples are not linearly separable.

Faculty Name : Mrs Swapna Subject Name :ML

Problem 1

Truth Table of OR Logical GATE is,

Weights w1 = 0.6, w2 = 0.6, Threshold = 1 and Learning Rate n = 0.5 are given

For Training Instance 1: A=0, B=0 and Target = 0

wi.xi = 0*0.6 + 0*0.6 = 0

This is not greater than the threshold of 1, so the output = 0. Here the target is same as calculated output.

For Training Instance 2: A=0, B=1 and Target = 1

wi.xi = 0*0.6 + 1*0.6 = 0.6

This is not greater than the threshold of 1, so the output = 0. Here the target does not match with calculated
output.

Now,

Weights w1 = 0.6, w2 = 1.1, Threshold = 1 and Learning Rate n = 0.5 are given

For Training Instance 1: A=0, B=0 and Target = 0

Faculty Name : Mrs Swapna Subject Name :ML

wi.xi = 0*0.6 + 0*1.1 = 0

This is not greater than the threshold of 1, so the output = 0. Here the target is same as calculated output.

For Training Instance 2: A=0, B=1 and Target = 1

wi.xi = 0*0.6 + 1*1.1 = 1.1

This is greater than the threshold of 1, so the output = 1. Here the target is same as calculated output.

For Training Instance 3: A=1, B=0 and Target = 1

wi.xi = 1*0.6 + 0*1.1 = 0.6

This is not greater than the threshold of 1, so the output = 0. Here the target does not match with calculated
output.

Now,

Weights w1 = 1.1, w2 = 1.1, Threshold = 1 and Learning Rate n = 0.5 are given

For Training Instance 1: A=0, B=0 and Target = 0

wi.xi = 0*2.2 + 0*1.1 = 0

This is not greater than the threshold of 1, so the output = 0. Here the target is same as calculated output.

For Training Instance 2: A=0, B=1 and Target = 1

wi.xi = 0*1.1 + 1*1.1 = 1.1

This is greater than the threshold of 1, so the output = 0. Here the target is same as calculated output.

For Training Instance 3: A=1, B=0 and Target = 1

Faculty Name : Mrs Swapna Subject Name :ML

wi.xi = 1*1.1 + 0*1.1 = 1.1

This is greater than the threshold of 1, so the output = 1. Here the target is same as calculated output.

For Training Instance 4: A=1, B=1 and Target = 1

wi.xi = 1*1.1 + 1*1.1 = 2.2

This is greater than the threshold of 1, so the output = 1. Here the target is same as calculated output.

Final wieghts w1 = 1.1, w2 = 1.1 Threshold = 1 and Learning Rate n = 0.5.

----------------------------------------------------------------------------------------------------------------------------------

Problem 2

Truth Table of AND Logical GATE is,

Weights w1 = 1.2, w2 = 0.6, Threshold = 1 and Learning Rate n = 0.5 are given

For Training Instance 1: A=0, B=0 and Target = 0

wi.xi = 0*1.2 + 0*0.6 = 0

This is not greater than the threshold of 1, so the output = 0, Here the target is same as calculated output.

Faculty Name : Mrs Swapna Subject Name :ML

For Training Instance 2: A=0, B=1 and Target = 0

wi.xi = 0*1.2 + 1*0.6 = 0.6

This is not greater than the threshold of 1, so the output = 0. Here the target is same as calculated output.

For Training Instance 2: A=1, B=0 and Target = 0

wi.xi = 1*1.2 + 0*0.6 = 1.2

This is greater than the threshold of 1, so the output = 1. Here the target does not match with the calculated
output.

Hence we need to update the weights.

Now,

After updating weights are w1 = 0.7, w2 = 0.6 Threshold = 1 and Learning Rate n = 0.5

W1 = 0.7, w2 = 0.6 Threshold = 1 and Learning Rate n = 0.5

For Training Instance 1: A=0, B=0 and Target = 0

wi.xi = 0*0.7 + 0*0.6 = 0

This is not greater than the threshold of 1, so the output = 0. Here the target is same as calculated output.

For Training Instance 2: A=0, B=1 and Target = 0

wi.xi = 0*0.7 + 1*0.6 = 0.6

This is not greater than the threshold of 1, so the output = 0. Here the target is same as calculated output.

For Training Instance 3: A=1, B=0 and Target = 0

Faculty Name : Mrs Swapna Subject Name :ML

wi.xi = 1*0.7 + 0*0.6 = 0.7

This is not greater than the threshold of 1, so the output = 0. Here the target is same as calculated output.

For Training Instance 4: A=1, B=1 and Target = 1

wi.xi = 1*0.7 + 1*0.6 = 1.3

This is greater than the threshold of 1, so the output = 1. Here the target is same as calculated output.

Hence the final weights are w1= 0.7 and w2 = 0.6, Threshold = 1 and Learning Rate n = 0.5.

------------------------------------------------------------------------------------------------------------------------------

An Example for NEURAL NETWORK REPRESENTATIONS in Real Time

A prototypical example of ANN learning is provided by author Pomerleau's system ALVINN,

(Autonomous Land Vehicle In a Neural Network) which uses a learned ANN to steer an
autonomous vehicle driving at normal speeds onpublic highways
The input to the neural network is a 30x32 grid of pixel intensities obtained from a forward-
pointed camera mounted on the vehicle.
The network output is the direction in which the vehicle is steered

Faculty Name : Mrs Swapna Subject Name :ML

 Figure: Neural network learning to steer an autonomous vehicle.
Figure illustrates the neural network representation.
The network is shown on the left side of the figure, with the input camera image depicted below it.
Each node (i.e., circle) in the network diagram corresponds to the output of a single network
unit, and the lines entering the node from below are its inputs.
There are four units that receive inputs directly from all of the 30 x 32 pixels in the image.
These are called "hidden" units because their output is available only within thenetwork and is
not available as part of the global network output. Each of these four hidden units computes a
single real-valued output based on a weighted combination of its 960 inputs
These hidden unit outputs are then used as inputs to a second layer of 30 "output" units.
Each output unit corresponds to a particular steering direction, and the output values of these
units determine which steering direction is recommended most strongly.
The diagrams on the right side of the figure depict the learned weight values associated with
one of the four hidden units in this ANN.
The large matrix of black and white boxes on the lower right depicts the weights from the 30
x 32 pixel inputs into the hidden unit. Here, a white box indicates a positive weight, a black
box a negative weight, and the size of the box indicates the weight magnitude.
--------------------------------------------------------------------------------------------------------

APPROPRIATE PROBLEMS FOR NEURAL NETWORK LEARNING

ANN learning is well-suited to problems in which the training data corresponds to noisy,
complex sensor data, such as inputs from cameras and microphones.

Faculty Name : Mrs Swapna Subject Name :ML

 ANN is appropriate for problems with the following characteristics:

1. Instances which are represented by many attribute-value pairs.

2. The target function output may be discrete-valued, real-valued, or a vector of several real-
or discrete-valued attributes.
3. The training examples may contain errors.
4. Long training times are acceptable to train.
5. Fast evaluation of the learned target function it can also support fast testing process.
Multi-layer Perceptron in TensorFlow

Multi-Layer perceptron defines the most complex architecture of artificial neural networks. It is
substantially formed from multiple layers of the perceptron. TensorFlow is a very popular deep learning
framework released by, and this notebook will guide to build a neural network with this library. If we want
to understand what is a Multi-layer perceptron, The pictorial representation of multi-layer perceptron
learning is as shown below-

MLP networks are used for supervised learning format. A typical learning algorithm for MLP networks is
also called back propagation's algorithm.

A multilayer perceptron (MLP) is a feed forward artificial neural network that generates a set of outputs
from a set of inputs. An MLP is characterized by several layers of input nodes connected as a directed graph
between the input nodes connected as a directed graph between the input and output layers. MLP uses back
propagation for training the network. MLP is a deep learning method.

MULTILAYER NETWORKS AND THE BACKPROPAGATION ALGORITHM example

Multilayer networks learned by the BACKPROPAGATION algorithm are capable of expressing a

rich variety of nonlinear decision surfaces.

Faculty Name : Mrs Swapna Subject Name :ML

Consider the example:

Here the speech recognition task involves distinguishing among 10 possible vowels, allspoken
in the context of "h_d" (i.e., "hid," "had," "head," "hood," etc.).
The network input consists of two parameters, F1 and F2, obtained from a spectral analysis of
the sound. The 10 network outputs correspond to the 10 possible vowel sounds. The network
prediction is the output whose value is highest.
The plot on the right illustrates the highly nonlinear decision surface represented by thelearned
network. Points shown on the plot are test examples distinct from the examples used to train
the network.

What is back propagation?

We can define the back propagation algorithm as an algorithm that trains some given feed-forward Neural
Network for a given input pattern where the classifications are known to us. At the point when every
passage of the example set is exhibited to the network, the network looks at its yield reaction to the
example input pattern. After that, the comparison done between output response and expected output with
the error value is measured. Later, we adjust the connection weight based upon the error value measured.

In simple terms, after each feed-forward passes through a network, this algorithm does the backward pass
to adjust the model’s parameters based on weights and biases. A typical supervised learning algorithm
attempts to find a function that maps input data to the right output. Back propagation works with a multi-
layered neural network and learns internal representations of input to output mapping.

How does back propagation work?

Let us take a look at how back propagation works. It has four layers: input layer, hidden layer, hidden
layer II and final output layer.

So, the main three layers are:

1. Input layer
2. Hidden layer

Faculty Name : Mrs Swapna Subject Name :ML

 3. Output layer
Each layer has its own way of working and its own way to take action such that we are able to get the
desired results and correlate these scenarios to our conditions. Let us discuss other details needed to help
summarizing this algorithm.

This image summarizes the functioning of the backpropagation approach.

1. Input layer receives x

2. Input is modeled using weights w
3. Each hidden layer calculates the output and data is ready at the output layer
4. Difference between actual output and desired output is known as the error
5. Go back to the hidden layers and adjust the weights so that this error is reduced in future runs
This process is repeated till we get the desired output. The training phase is done with supervision. Once
the model is stable, it is used in production.

Why do we need back propagation?

Back propagation has many advantages, some of the important ones are listed below-

 Back propagation is fast, simple and easy to implement

 There are no parameters to be tuned
 Prior knowledge about the network is not needed thus becoming a flexible method
 This approach works very well in most cases

Feed forward network

Feed forward networks are also called MLN i.e Multi-layered Networks. They are known as feed-forward
because the data only travels forward in NN through input node, hidden layer and finally to the output
nodes. It is the simplest type of artificial neural network.

Disadvantages of using Backpropagation

 The actual performance of backpropagation on a specific problem is dependent on the input data.
 Back propagation algorithm can be quite sensitive to noisy data

Faculty Name : Mrs Swapna Subject Name :ML

Gradient Descent
A gradient measures how much the output of a function changes if you change the inputs a little bit."
— Lex Fridman (MIT)

The key idea behind the delta rule is to use gradient descent to search the hypothesis space
of possible weight vectors to find the weights that best fit the training examples.
To understand the delta training rule, consider the task of training a threshold perception. That is, a
linear unit for which the output O is given by

To derive a weight learning rule for linear units, specify a measure for the training error of a
hypothesis (weight vector), relative to the training examples.

Where,
D is the set of training examples,
td is the target output for training example d,
od is the output of the linear unit for training example d
E ( w→ ) is simply half the squared difference between the target output td and the linear unit
output od, summed over all training examples.

Gradient Descent Error Estimation in 3 dimensional plane

A gradient simply measures the change in all weights with regard to the change in error. You can also think
of a gradient as the slope of a function. The higher the gradient, the steeper the slope and the faster a model
can learn. But if the slope is zero, the model stops learning.

Visualizing the Hypothesis Space

To understand the gradient descent algorithm, it is helpful to visualize the entire hypothesis
space of possible weight vectors and their associated E values as shown in below figure.
Here the axes w0 and wl represent possible values for the two weights of a simple linear unit.
The w0, wl plane therefore represents the entire hypothesis space.
The vertical axis indicates the error E relative to some fixed set of training examples.
The arrow shows the negated gradient at one particular point, indicating the direction inthe w0,
wl plane producing steepest descent along the error surface.
The error surface shown in the figure thus summarizes the desirability of every weight vector
in the hypothesis space

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
 Given the way in which we chose to define E, for linear units this error surface mustalways
be parabolic with a single global minimum.

Gradient descent search determines a weight vector that minimizes E by starting with an arbitrary
initial weight vector, then repeatedly modifying it in small steps.
At each step, the weight vector is altered in the direction that produces the steepest descent along the
error surface depicted in above figure. This process continues until the global minimum error is
reached.

Types of Gradient Descent

There are three popular types of gradient descent that mainly differ in the amount of data they use:
BATCH GRADIENT DESCENT
Batch gradient descent, also called vanilla gradient descent, calculates the error for each example within the
training dataset, but only after all training examples have been evaluated does the model get updated. This
whole process is like a cycle and it's called a training epoch.
Some advantages of batch gradient descent are its computational efficient, it produces a stable error gradient
and a stable convergence. Some disadvantages are the stable error gradient can sometimes result in a state
of convergence that isn’t the best the model can achieve. It also requires the entire training dataset be in
memory and available to the algorithm.
STOCHASTIC GRADIENT DESCENT
By contrast, stochastic gradient descent (SGD) does this for each training example within the dataset,
meaning it updates the parameters for whole training data set example one by one. Depending on the
problem, this can make SGD faster than batch gradient descent.
The frequent updates, however, are more computationally expensive than the batch gradient descent
approach.

Faculty Name : Mrs Swapna Subject Name :ML

Additionally, the frequency of those updates can result in noisy gradients, which may cause the error rate to
jump around instead of slowly decreasing.
MINI-BATCH GRADIENT DESCENT
Mini-batch gradient descent is the go-to method since it’s a combination of the concepts of SGD and batch
gradient descent. It simply splits the training dataset into small batches and performs an update for each of
those batches.
This creates a balance between the robustness of stochastic gradient descent and the efficiency of batch
gradient descent.
--------------------------------------------------------------------------------------------------------------------------------
Derivation of the Gradient Descent Rule

How to calculate the direction of steepest descent along the error surface?

The direction of steepest can be found by computing the derivative of E with respect to each
component of the vector w→ . This vector derivative is called the gradient of E with respect to
w→ , written as

The gradient specifies the direction of steepest increase of E, the training rule for gradient
descent is

Here η is a positive constant called the learning rate, which determines the step size in the
gradient descent search.
The negative sign is present because we want to move the weight vector in the direction
that decreases E.

This training rule can also be written in its component form

Faculty Name : Mrs Swapna Subject Name :ML

Calculate the gradient at each step. The vector derivatives that form the
gradient can be obtained by differentiating E from Equation (2), as

Faculty Name : Mrs Swapna Subject Name :ML

 To summarize, the gradient descent algorithm for training linear units is as follows:

Pick an initial random weight vector.

Apply the linear unit to all training examples, then compute Δwi for each weight
according to

Update each weight wi by adding Δwi, then repeat this process

Issues in Gradient Descent Algorithm

 Can veer off in the wrong direction due to frequent updates.

 Frequent updates are computationally expensive in process due to using all resources for
processing one training sample at a time.

Example of Feed Forward Network and back propagation in Real time: Face recognition

Face recognition using neural network explains about concept of improving performance of
detecting face by using neural technology. Fundamental part of face recognition is done through

Faculty Name : Mrs Swapna Subject Name :ML

face detection system. Problems with face detection from arbitrary images are due to changes in
skin color, quality of image position and orientation.

Different set of multilayer neural network work on detection of face and back propagation
algorithm is used for error detection when the face is undetected by machine.

Faculty Name : Mrs Swapna Subject Name :ML

 Face Recognition Tasks
The task of face recognition is broad and can be tailored to the specific needs of a prediction
problem.

For example, in the 1995 paper titled “Human and machine recognition of faces: A survey,” the
authors describe three face recognition tasks:
 Face Matching: Find the best match for a given face.
 Face Similarity: Find faces that are most similar to a given face.
 Face Transformation: Generate new faces that are similar to a given face.

With face detection, you can get the information you need to perform tasks like embellishing
selfies and portraits, or generating avatars from a user's photo.

Faculty Name : Mrs Swapna Subject Name :ML

 Key capabilities

 Recognize and locate facial features Get the coordinates of the eyes, ears, cheeks, nose, and
mouth of every face detected.
 Get the contours of facial features Get the contours of detected faces and their eyes, eyebrows,
lips, and nose.
 Recognize facial expressions Determine whether a person is smiling or has their eyes closed.
 Track faces across video frames Get an identifier for each unique detected face. The identifier
is consistent across invocations, so you can perform image manipulation on a particular person in
a video stream.
 Process video frames in real time Face detection is performed on the device, and is fast enough
to be used in real-time applications, such as video manipulation.
Face alignment
Face alignment is a computer vision technology for identifying the geometric structure of
human faces in digital images. Given the location and size of a face, it automatically
determines the shape of the face components such as eyes and nose.

Feature extraction refers to the process of transforming raw data into numerical features
that can be processed while preserving the information in the original data set.

Feature matching refers to finding corresponding features from two similar images based on
a search distance algorithm. One of the images is considered the source and the other as target,
and the feature matching technique is used to either find or derive and transfer attributes from
source to target image
---------------------------------------------------------------------------------------------------------------------

Advance Topics in neural network

1. Alternative Error Functions
2. Alternative Error Minimization Procedures
3. Recurrent Networks
4. Dynamically Modifying Network Structure

1.Alternative Error Functions

the basic BACKPROPAGATION algorithm defines E in terms of the sum of squared errors
of the network, other definitions have been suggested in order to incorporate other constraints

Faculty Name : Mrs Swapna Subject Name :ML

 into the weight-tuning rule. For each new definition of E a new weight-tuning rule for
gradient descent must be derived. Examples of alternative definitions of E include a Adding a
penalty term for weight magnitude to Adjust Weights to achieve Global minima point

the new penalty weight added is

Gamma is constant term and wji is new penalty weight adjusted for error reduction

2. Alternative Error Minimization Procedures

1.Weight-update method

Direction: choosing a direction in which to alter or converge the current weight vector (ex:
the gradient in Backpropagation) which depends on Distance : choosing a distance to move
(ex: the learning ratio η )

Ex : Line search method, Conjugate gradient method

Line search method is an iterative approach to find a local minimum of a multidimensional

nonlinear function using the function's gradients. It computes a search direction and then finds an
acceptable step length that Line search method can be categorized into exact and inexact
methods.

Gradient descent is computationally efficient, provides a slow rate of convergence. This is where
line search comes into place and provides much better rate of convergence at a slight
increase in computational and accuracy

The conjugate gradient method is a mathematical technique that can be useful for the
optimization of both linear and non-linear systems. This technique is generally used as an
iterative algorithm, however, it can be used as a direct method, and it will produce a numerical
solution. Generally this method is used for very large systems where it is not practical to solve
with a direct method of Gradient Decent back propagation.

3.Recurrent Networks

A recurrent neural network (RNN) is a type of artificial neural network which uses sequential
data or time series data which are dynamic in nature. These deep learning algorithms are
commonly used for ordinal or temporal problems, such as language translation, natural language

Faculty Name : Mrs Swapna Subject Name :ML

 processing (nlp), speech recognition, and image captioning; they are incorporated into popular
applications such as Siri, voice search, and Google Translate.

Like feedforward and convolutional neural networks (CNNs), recurrent neural networks utilize
training data to learn. They are distinguished by their “memory” as they take information from
prior inputs to influence the current input and output within a specific time step .

Types of neural Network

1. One to many network

2. One to one network
3. Many to one network
4. Many to many network

4. Dynamically Modified network structure

Dynamic Neural networks can be considered as the improvement of the static neural networks in
which by adding more decision algorithms we can make neural networks learning dynamically
from the input and generate better quality results.

Modifying the network structure in hidden layer by adding and pruning the structure for better
results and accuracy.

-------------------------------------------------------------------------------------------------------------------

REMARKS ON THE BACKPROPAGATION ALGORITHM

1. Convergence to Local Minima Global Mininum

 The BACKPROPAGATION multilayer networks is only guaranteed to converge toward
some local minimum in E and not necessarily to the global minimum error.
 Despite the lack of assured convergence to the global minimum error,
BACKPROPAGATION is a highly effective function approximation method in practice.
 To alleviate this convergence we have frequently update the weights
 Can use batch and stochastic gradient decent to improve the efficiency and reduction in
error is possible in multilayer network.
2. Representational Power of Feedforward Networks

What set of functions can be represented by feed-forward networks?

Faculty Name : Mrs Swapna Subject Name :ML

 The answer depends on the width and depth of the networks. There are three quite
general results are known about which function classes can be described by which types
of Networks

1. Boolean functions – Every boolean function can be represented exactly by some network
with two layers of units, although the number of hidden units required grows exponentially
in the worst case with the number of network inputs

2. Continuous functions – Every bounded continuous function can be approximated with

arbitrarily small error by a network with two layers of units

3. Arbitrary functions – Any function can be approximated to arbitrary accuracy by a

network with three layers of units.

3. Hypothesis Space Search and Inductive Bias

 Hypothesis space is the n-dimensional Euclidean space of the n network weights and
hypothesis space is continuous.
 As it is continuous, E is differentiable with respect to the continuous parameters of the
hypothesis, results in a well-defined error gradient that provides a very useful structure for
organizing the search for the best hypothesis.

 It is difficult to characterize precisely the inductive bias of BACKPROPAGATION

algorithm, because it depends on the interplay between the gradient descent and the way in
which the weight space are adjusted to achieve global and local minima.

 However, one can roughly characterize it as smooth interpolation between different data
nodes between input ,output and hiddenlayer .

4. Hidden Layer Representations

BACKPROPAGATION can define new hidden layer features that are not explicit in the
input representation, but which capture properties of the input instances that are most
relevant to learning the target function.

Consider example, the network shown in below Figure

Faculty Name : Mrs Swapna Subject Name :ML

  Consider training the network shown in Figure to learn the simple target function f (x)
= x, where x is a vector containing seven 0's and a single 1.
 The network must learn to reproduce the eight inputs at the corresponding eight
output units. Although this is a simple function, the network in this case is
constrained to use only three hidden units. Therefore, the essential information
from all eight input units must be captured by the three learned hidden units.
 When BACKPROPAGATION applied to this task, using each of the eight possible
vectors as training examples, it successfully learns the target function. By
examining the hidden unit values generated by the learned network for each of the
eight possible input vectors, it is easy to see that the learned encoding is similar to

Faculty Name : Mrs Swapna Subject Name :ML

 the familiar standard binary encoding of eight values using three bits (e.g.,
000,001,010,. . . , 111). The exact values of the hidden units for one typical run of
shown in Figure.
 This ability of multilayer networks to automatically discover useful representations
at the hidden layers is a key feature of ANN learning

5. Generalization of weights reduces errors, Over fitting data makes the error reduction
process complex, and Stopping Criterion when error is optimal when achieved global
minima point.

What is an appropriate condition for terminating the weight update loop?

One choice is to continue training until the error E on the training examples falls below
some predeterminedthreshold.
To see the dangers of minimizing the error over the training data, consider how the
error Evaries with the number of weight iterations

Faculty Name : Mrs Swapna Subject Name :ML

  Consider first the top plot in this figure. The lower of the two lines shows the
monotonically decreasing error E over the training set, as the number of gradient
descent iterations grows.
 The upper line shows the error E measured over a different validation set of
examples, distinct from the training examples. This line measures the
generalization accuracy of the network-the accuracy with which it fits examples
beyondthe training data.

 The generalization accuracy measured over the validation examples first decreases,
thenincreases, even as the error over the training examples continues to decrease.

When the training data set increases over fitting increase the complexity of error
and reduction of the error may take n number of iterations to modify the weights
which becomes a tedious task.

Estimating Hypothesis Accuracy

A model is constructed based on hypothesis and estimating hypothesis based on accuracy which
is best this is based on sample data or additional sample data which is taken into consideration in
learning.

Different instances in the model are taken into consideration

Faculty Name : Mrs Swapna Subject Name :ML

 Example in h1 hypothesis is having 100 instances and h2 hypothesis have 50 instances

The estimation is based on sample data and error frequency in different instances

Acccuracy is based on Error Levels in the model

This is made clear by distinguishing between the true error of a model and the estimated or
sample error.

 Sample Error. Estimate of error calculated on a sample data.

 The sample error (errors(h)) of hypothesis h with respect to target function f and data
sample S is

Where n is the number of examples in S, and the quantity δ(f(x), h(x)) is 1 if error is
identified

if f (x) ≠ h(x), and 0 no error identified.

 True Error: Estimation of Error over entire distribution

 The true error (errorD(h)) of hypothesis h with respect to target function f and distribution
D, is the probability that h will misclassify an instance drawn at random accordingto D.

Confidence Intervals for Discrete-Valued Hypotheses

Suppose we wish to estimate the true error for some discrete valued hypothesis h, based on its
observed sample error over a sample S, where
 The sample S contains n examples drawn independent of one another, and independent
of h, according to the probability distribution D
 n ≥ 30
 Hypothesis h commits r errors over these n examples (i.e., errors (h) = r/n).

Faculty Name : Mrs Swapna Subject Name :ML

 Under these conditions, statistical theory allows to make the following assertions:
1. Given no other information, the most probable value of errorD (h) is errors(h)
2. With approximately 95% probability, the true error errorD (h) lies in the interval

Example:

Suppose the data sample S contains n = 40 examples and that hypothesis h commits r =
12 errors over this data.
 The sample error is errors(h) = r/n = 12/40 = 0.30
 Given no other information, true error is errorD (h) = errors(h), i.e., errorD (h)
=0.30
 With the 95% confidence interval estimate for errorD (h).

= 0.30 ± (1.96 * 0.07) = 0.30 ± 0.14

3. A different constant, ZN, is used to calculate the N% confidence interval. The general
expression for approximate N% confidence intervals for errorD (h) is

Where,

Faculty Name : Mrs Swapna Subject Name :ML

 The above equation describes how to calculate the confidence intervals, or error bars,
for estimates of errorD (h) that are based on errors(h)

Example:

Suppose the data sample S contains n = 40 examples and that hypothesis h commits r =
12 errors over this data.
 The sample error is errors(h) = r/n = 12/40 = 0.30
 With the 68% confidence interval estimate for errorD (h).

= 0.30 ± (1.00 * 0.07)

= 0.30 ± 0.07

Basics of Sampling Theory

Sampling theory is the field of statistics that is involved
with the collection, analysis and interpretation of data gathered
from random samples of a population under study.

The application of sampling theory is concerned not only with the proper
selection of observations from the population that will
constitute the random sample; it also involves the use of
probability theory, along with prior knowledge about the
population parameters, to analyze the data from the random sample
and develop conclusions from the analysis

Faculty Name : Mrs Swapna Subject Name :ML

The Binomial Distribution

Consider the following problem for better understanding of Binomial Distribution

 Given a worn and bent coin and estimate the probability that the coin will turn up heads
when tossed.
 Unknown probability of heads p. Toss the coin n times and record the number of times
r that it turns up heads.
Estimate of p = r / n

 If the experiment were rerun, generating a new set of n coin tosses, we might expect the
number of heads r to vary somewhat from the value measured in the first experiment,
yielding a somewhat different estimate for p.
 The Binomial distribution describes for each possible value of r (i.e., from 0 to n), the
probability of observing exactly r heads given a sample of n independent tosses of a

Faculty Name : Mrs Swapna Subject Name :ML

 coin whose true probability of heads is p.
COMPARING LEARNING ALGORITHMS

Which are important parameter of hypothesis testing ?

Null hypothesis :- In inferential statistics(make predictions (“inferences”) from that data.), the
null hypothesis is a general statement or default position that there is no relationship between two
measured phenomena, or no association among groups
In other words it is a basic assumption or made based on domain or problem knowledge.
Ex : a company production is = 50 unit/per day etc.

Alternative hypothesis :-
The alternative hypothesis is the hypothesis used in hypothesis testing that is contrary to the null
hypothesis. It is usually taken to be that the observations are the result of a real effect (with some
evidence)
Level of significance: Refers to the degree of significance in which we accept or reject the null-
hypothesis. 100% accuracy is not possible for accepting or rejecting a hypothesis, so we therefore
select a level of significance that is usually 5%.
This is normally denoted with alpha(maths symbol ) and generally it is 0.05 or 5% , which means
your output should be 95% confident to give similar kind of result in each sample.

Faculty Name : Mrs Swapna Subject Name :ML

 Type I error: When we reject the null hypothesis, although that hypothesis was true. Type I error
is denoted by alpha. In hypothesis testing, the normal curve that shows the critical region is called
the alpha region
Type II errors: When we accept the null hypothesis but it is false. Type II errors are denoted by
beta. In Hypothesis testing, the normal curve that shows the acceptance region is called the beta
region.

One tailed test :- A test of a statistical hypothesis , where the region of rejection is on
only one side of the sampling distribution , is called a one-tailed test.
Two-tailed test :- A two-tailed test is a statistical test in which the critical area of a distribution
is two-sided and tests whether a sample is greater than or less than a certain range of values. If the
sample being tested falls into either of the critical areas, the alternative hypothesis is accepted
instead of the null hypothesis.
Some of widely used hypothesis testing type( not in syllabus)
1. T Test

Faculty Name : Mrs Swapna Subject Name :ML

2. Z Test
3. F- Test
4. ANOVA
5. Chi-Square Test

Faculty Name : Mrs Swapna Subject Name :ML

 UNIT 3

Bayes Theorem provides a principled way for calculating a conditional probability.

Bayes Theorem is also widely used in the field of machine learning. Including its use in a
probability framework for fitting a model to a training dataset, referred to as maximum a
posteriori or MAP for short, and in developing models for classification predictive modeling
problems such as the Bayes Optimal Classifier and Naive Bayes.

 Joint Probability: Probability of two (or more) simultaneous events, e.g. P(A and B) or P(A, B).
The conditional probability is the probability of one event given the occurrence of another event,
often described in terms of events A and B from two dependent random variables e.g. X and Y.

 Conditional Probability: Probability of one (or more) event given the occurrence of another
event, e.g. P(A given B) or P(A | B).
The joint probability can be calculated using the conditional probability; for example:

 P(A, B) = P(A | B) * P(B)

This is called the product rule. Importantly, the joint probability is symmetrical, meaning that:

 P(A, B) = P(B, A)
The conditional probability can be calculated using the joint probability; for example:

 P(A | B) = P(A, B) / P(B)

An Alternate Way To Calculate Conditional Probability
The conditional probability can be calculated using the other conditional probability; for
example:

 P(A|B) = P(B|A) * P(A) / P(B)

The reverse is also true; for example:

 P(B|A) = P(A|B) * P(B) / P(A)

Bayes theorem is a theorem in probability and statistics, named after the Reverend Thomas
Bayes, that helps in determining the probability of an event that is based on some event that has
already occurred. Bayes theorem has many applications such as bayesian interference, in the
healthcare sector - to determine the chances of developing health problems with an increase in
age and many others.

Faculty Name : Mrs Swapna Subject Name :ML

 Bayes theorem, in simple words, determines the conditional probability of an event A given that
event B has already occurred. Bayes theorem is also known as the Bayes Rule or Bayes Law.

It can be helpful to think about the calculation from these different perspectives and help to map
your problem onto the equation.

Firstly, in general, the result P(A|B) is referred to as the posterior probability and P(A) is
referred to as the prior probability.
 P(A|B): Posterior probability.
 P(A): Prior probability.
Sometimes P(B|A) is referred to as the likelihood and P(B) is referred to as the evidence.
 P(B|A): Likelihood.
 P(B): Evidence.
This allows Bayes Theorem to be restated as:

 Posterior = Likelihood * Prior / Evidence

Maximum a Posteriori (MAP) Hypothesis

 In many learning scenarios, the learner considers some set of candidate

hypotheses H and is interested in finding the most probable hypothesis h ∈
H given the observed data

Faculty Name : Mrs Swapna Subject Name :ML

 D. Any such maximally probable hypothesis is called a maximum a posteriori
(MAP) hypothesis.
 Bayes theorem to calculate the posterior probability of each candidate
hypothesis is hMAP is a MAP hypothesis provided

 P(D) can be dropped, because it is a constant independent of h

Notations
 P(h) prior probability of h, reflects any background knowledge about the
chance that h is correct
 P(D) prior probability of D, probability that D will be observed
 P(D|h) probability of observing D given a world in which h holds
 P(h|D) posterior probability of h, reflects confidence that h holds after D
has been observed

Maximum Likelihood (ML) Hypothesis

 In some cases, it is assumed that every hypothesis in H is equally

probable a priori (P(hi) = P(hj) for all hi and hj in H).
 In this case the below equation can be simplified and need only consider the
term P(D|h) to find the most probable hypothesis.

Faculty Name : Mrs Swapna Subject Name :ML

 P(D|h) is often called the likelihood of the data D given h, and any hypothesis that
maximizes P(D|h) is called a maximum likelihood (ML) hypothesis

Example
 Consider a medical diagnosis problem in which there are two alternative
hypotheses:
(1) that the patient has particular form of cancer, and (2) that the patient
does not. The available data is from a particular laboratory test with two
possible outcomes: + (positive) and - (negative).

 We have prior knowledge that over the entire population of people only
.008 have this disease. Furthermore, the lab test is only an imperfect
indicator of the disease.
 The test returns a correct positive result in only 98% of the cases in which
the disease is actually present and a correct negative result in only 97% of
the cases in which the disease is not present. In other cases, the test
returns the opposite result.
 The above situation can be summarized by the following probabilities:

Suppose a new patient is observed for whom the lab test returns a positive
(+) result. Should we diagnose the patient as having cancer or not?

The exact posterior probabilities can also be determined by normalizing the above
quantities so that they sum

Faculty Name : Mrs Swapna Subject Name :ML

Result : The Patient is not Having Cancer

BAYES THEOREM AND CONCEPT LEARNING

What is the relationship between Bayes theorem and the problem of concept learning?

Since Bayes theorem provides a principled way to calculate the posterior

probability of each hypothesis given the training data, and can use it as the basis for
a straightforward learning algorithm that calculates the probability for each
possible hypothesis, then outputs the most probable.

Brute-Force Bayes Concept Learning

Consider the concept learning problem

 Assume the learner considers some finite hypothesis space H defined over
the instance space X, in which the task is to learn some target concept c : X
→ {0,1}.
 Learner is given some sequence of training examples ((x1, d1) . . . (xm, dm))
where xi is some instance from X and where di is the target value of xi (i.e.,
di = c(xi)).
 The sequence of target values are written as D = (d1 . . . dm).

We can design a straightforward concept learning algorithm to output the maximum

a posteriori hypothesis, based on Bayes theorem, as follows:

BRUTE-FORCE MAP LEARNING algorithm:

1. For each hypothesis h in H, calculate the posterior probability

2. Output the hypothesis hMAP with the highest posterior probability

Faculty Name : Mrs Swapna Subject Name :ML

 In order specify a learning problem for the BRUTE-FORCE MAP LEARNING
algorithm we must specify what values are to be used for P(h) and for P(D|h) ?

Let’s choose P(h) and for P(D|h) to be consistent with the following assumptions:
 The training data D is noise free (i.e., di = c(xi))
 The target concept c is contained in the hypothesis space H
 Do not have a priori reason to believe that any hypothesis is more probable
than any other.
What values should we specify for P(h)?

 Given no prior knowledge that one hypothesis is more likely than another,
it is reasonable to assign the same prior probability to every hypothesis h in
H.
 Assume the target concept is contained in H and require that these prior
probabilities sum to 1.

What choice shall we make for P(D|h)?

 P(D|h) is the probability of observing the target values D = (d1 . . .dm) for
the fixed set of instances (x1 . . . xm), given a world in which hypothesis h
holds
 Since we assume noise-free training data, the probability of observing
classification di given h is just 1 if di = h(xi) and 0 if di ≠ h(xi). Therefore,

Given these choices for P(h) and for P(D|h) we now have a fully-defined problem
for the above BRUTE-FORCE MAP LEARNING algorithm.

Recalling Bayes theorem, we have

Faculty Name : Mrs Swapna Subject Name :ML

 Consider the case where h is inconsistent with the training data D

The posterior probability of a hypothesis inconsistent with D is zero

Consider the case where h is consistent with D

Where, VSH,D is the subset of hypotheses from H that are consistent with D

To summarize, Bayes theorem implies that the posterior probability P(h|D) under our
assumed P(h) and P(D|h) is

MAXIMUM LIKELIHOOD AND LEAST-SQUARED ERROR HYPOTHESES

Consider the problem of learning a continuous-valued target function such as

neural network learning, linear regression, and polynomial curve fitting

A straightforward Bayesian analysis will show that under certain assumptions any
learning algorithm that minimizes the squared error between the output hypothesis
predictions and the training data will output a maximum likelihood (ML) hypothesis

Faculty Name : Mrs Swapna Subject Name :ML

  Learner L considers an instance space X and a hypothesis space H
consisting of some class of real-valued functions defined over X, i.e., (∀ h ∈
H)* h : X → R+ and training examples of the form <xi,di>
 The problem faced by L is to learn an unknown target function f : X → R
 A set of m training examples is provided, where the target value of each
example is corrupted by random noise drawn according to a Normal
probability distribution with zero mean (di = f(xi) + ei)
 Each training example is a pair of the form (xi ,di ) where di = f (xi ) + ei .
– Here f(xi) is the noise-free value of the target function and ei is a
random variable representing the noise.
– It is assumed that the values of the ei are drawn independently and
that they are distributed according to a Normal distribution with
zero mean.
 The task of the learner is to output a maximum likelihood hypothesis or a
MAP hypothesis assuming all hypotheses are equally probable a priori.

Using the definition of hML we have

Assuming training examples are mutually independent given h, we can write

P(D|h) as the product of the various (di|h)

Given the noise ei obeys a Normal distribution with zero mean and unknown
variance ζ2 , each di must also obey a Normal distribution around the true
targetvalue f(xi). Because we are writing the expression for P(D|h), we assume h is
the correct description of f.

Faculty Name : Mrs Swapna Subject Name :ML

 Hence, µ = f(xi) = h(xi)

Maximize the less complicated logarithm, which is justified because of the monotonicity
of function p

The first term in this expression is a constant independent of h, and can

therefore be discarded, yielding

Maximizing this negative quantity is equivalent to minimizing the

corresponding positive quantity

Faculty Name : Mrs Swapna Subject Name :ML

 Finally, discard constants that are independent of h.

Thus, above equation shows that the maximum likelihood hypothesis hML is the one
that minimizes the sum of the squared errors between the observed training values
di and the hypothesis predictions h(xi)

MAXIMUM LIKELIHOOD HYPOTHESES FOR PREDICTING PROBABILITIES

 Consider the setting in which we wish to learn a nondeterministic

(probabilistic) function f : X → ,0, 1-, which has two discrete output values.
 We want a function approximator whose output is the probability that f(x)
= 1. In other words, learn the target function f ` : X → *0, 1+ such that f ` (x)
= P(f(x) = 1)

How can we learn f ` using a neural network?

 Use of brute force way would be to first collect the observed frequencies of
1's and 0's for each possible value of x and to then train the neural network
to output the target frequency for each x.

What criterion should we optimize in

order to find a maximum likelihood hypothesis for f' in this setting?

 First obtain an expression for P(D|h)

 Assume the training data D is of the form D = {(x1, d1) . . . (xm, dm)}, where di
is the observed 0 or 1 value for f (xi).
 Both xi and di as random variables, and assuming that each training
example is drawn independently, we can write P(D|h) as

Faculty Name : Mrs Swapna Subject Name :ML

 Applying
the product rule

The probability P(di|h, xi)

Re-express it in a more mathematically manipulable form, as

Equation (4) to substitute for P(di |h, xi) in Equation (5) to obtain

We write an expression for the maximum likelihood hypothesis

The last term is a constant independent of h, so it can be dropped

It easier to work with the log of the likelihood, yielding

Equation (7) describes the quantity that must be maximized in order to obtain the
maximum likelihood hypothesis in our current problem setting

Faculty Name : Mrs Swapna Subject Name :ML

The Evolution of Probabilities Associated with Hypotheses

 Figure (a) all hypotheses have the same probability.

 Figures (b) and (c), As training data accumulates, the posterior
probability for inconsistent hypotheses becomes zero while the total
probability summing to 1 is shared equally among the remaining
consistent hypotheses.

MAP Hypotheses and Consistent Learners

 A learning algorithm is a consistent learner if it outputs a hypothesis that

commits zero errors over the training examples.
 Every consistent learner outputs a MAP hypothesis, if we assume a uniform
prior probability distribution over H (P(hi) = P(hj) for all i, j), and
deterministic, noise free training data (P(D|h) =1 if D and h are consistent,
and 0 otherwise).

Example:
 FIND-S outputs a consistent hypothesis, it will output a MAP hypothesis
under the probability distributions P(h) and P(D|h) defined above.
 Are there other probability distributions for P(h) and P(D|h) under which
FIND-S outputs MAP hypotheses? Yes.
Faculty Name : Mrs Swapna Subject Name :ML
  Because FIND-S outputs a maximally specific hypothesis from the version
space, its output hypothesis will be a MAP hypothesis relative to any prior
probability distribution that favors more specific hypotheses.

Note

 Bayesian framework is a way to characterize the behavior of learning algorithms

 By identifying probability distributions P(h) and P(D|h) under which the
output is a optimal hypothesis, implicit assumptions of the algorithm can
be characterized (Inductive Bias)

MINIMUM DESCRIPTION LENGTH PRINCIPLE

 A Bayesian perspective on Occam’s razor

 Motivated by interpreting the definition of hMAP in the light of basic
concepts from information theory.

which can be equivalently expressed in terms of maximizing the log2

or alternatively, minimizing the negative of this quantity

This equation (1) can be interpreted as a statement that short hypotheses are
preferred, assuming a particular representation scheme for encoding hypotheses
and data

 -log2P(h): the description length of h under the optimal encoding for the
hypothesis space H, LCH (h) = −log2P(h), where CH is the optimal code for
hypothesis space H.
 -log2P(D | h): the description length of the training data D given hypothesis

Faculty Name : Mrs Swapna Subject Name :ML

 h, under the optimal encoding from the hypothesis space H: LCH (D|h) =
−log2P(D| h) , where C D|h is the optimal code for describing data D
assuming that both the sender and receiver know the hypothesis h.
 Rewrite Equation (1) to show that hMAP is the hypothesis h that minimizes
the sum given by the description length of the hypothesis plus the
description length of the data given the hypothesis.

Where, CH and CD|h are the optimal encodings for H and for D given h

The Minimum Description Length (MDL) principle recommends choosing the

hypothesis that minimizes the sum of these two description lengths of equ.

Minimum Description Length principle:

Where, codes C1 and C2 to represent the hypothesis and the data given the hypothesis

The above analysis shows that if we choose C1 to be the optimal encoding of

hypotheses CH, and if we choose C2 to be the optimal encoding CD|h, then
hMDL = hMAP

Bayes Optimal Classifier

It is described using the Bayes Theorem that provides a principled way for calculating a
conditional probability. It is also closely related to the Maximum a Posteriori: a probabilistic
framework referred to as MAP that finds the most probable hypothesis for a training dataset.

In practice, the Bayes Optimal Classifier is computationally expensive, if not intractable to

calculate, and instead, simplifications such as the Gibbs algorithm and Naive Bayes can be used
to approximate the outcome.

Faculty Name : Mrs Swapna Subject Name :ML

 Bayes Theorem provides a principled way for calculating conditional probabilities, called a
posterior probability.
 Maximum a Posteriori is a probabilistic framework that finds the most probable hypothesis that
describes the training dataset.

 Bayes Optimal Classifier is a probabilistic model that finds the most probable prediction using
the training data and space of hypotheses to make a prediction for a new data instance.

To develop some intuitions consider a hypothesis space containing three hypotheses, hl, h2, and
h3. Suppose that the posterior probabilities of these hypotheses given the training data are .4, .3,
and .3 respectively. Thus, hl is the MAP hypothesis. Suppose a new instance x is encountered,
which is classified positive by hl, but negative by h2 and h3.

Taking all hypotheses into account, the probability that x is positive is .4 (the probability
associated with hi), and the probability that it is negative is therefore .6.

The most probable classification (negative) in this case is different from the classification
generated by the MAP hypothesis. In general, the most probable classification of the new
instance is obtained by combining the predictions of all hypotheses, weighted by their posterior
probabilities.

If the possible classification of the new example can take on any value vj from some set V, then
the probability P(vjlD) that the correct classification for the new instance is v;, is just

Faculty Name : Mrs Swapna Subject Name :ML

Two of the most commonly used simplifications use a sampling algorithm for
hypotheses, such as Gibbs sampling, or to use the simplifying assumptions of the Naive
Bayes classifier.
1.Gibbs Algorithm. Randomly sample hypotheses biased on their posterior probability.

2.Naive Bayes. Assume that variables in the input data are conditionally independent.

1.Gibbs Algorithm

Gibbs sampling (also called alternating conditional sampling) is a Markov Chain Monte
Carlo algorithm for high-dimensional data such as image processing and micro arrays.

Faculty Name : Mrs Swapna Subject Name :ML

It is called Monte Carlo because it draws samples from specified probability
distributions the Markov chain comes from the fact that each sample is dependent on the
previous sample. Gibbs sampling is relatively easy to implement. However, it is less
efficient than direct simulation from the distribution.

An alternative, less optimal method is the Gibbs algorithm defined as follows:

1. Choose a hypothesis h from H at random, according to the posterior probability
distribution over H.
2. Use h to predict the classification of the next instance x.
Given a new instance to classify, the Gibbs algorithm simply applies a hypothesis drawn
at random according to the current posterior probability distribution. Surprisingly, it can
be shown that under certain conditions the expected misclassification error for the
Gibbs algorithm is at most twice the expected error of the Bayes optimal classifier

2.Naive Bayes:

Assume that variables in the input data are conditionally independent.

o Naïve Bayes algorithm is a supervised learning algorithm, which is based on Bayes
theorem and used for solving classification problems.
o It is mainly used in text classification that includes a high-dimensional training dataset.
o Naïve Bayes Classifier is one of the simple and most effective Classification algorithms
which helps in building the fast machine learning models that can make quick
predictions.
o It is a probabilistic classifier, which means it predicts on the basis of the probability
of an object.
o Some popular examples of Naïve Bayes Algorithm are spam filtration, Sentimental
analysis, and classifying articles.

Working of Naïve Bayes' Classifier:

Working of Naïve Bayes' Classifier can be understood with the help of the below example:

Faculty Name : Mrs Swapna Subject Name :ML

Suppose we have a dataset of weather conditions and corresponding target variable "Play".
So using this dataset we need to decide that whether we should play or not on a particular day
according to the weather conditions. So to solve this problem, we need to follow the below steps:

1. Convert the given dataset into frequency tables.

2. Generate Likelihood table by finding the probabilities of given features.
3. Now, use Bayes theorem to calculate the posterior probability.

Problem: If the weather is sunny, then the Player should play or not?

Solution: To solve this, first consider the below dataset:

Outlook Play

0 Rainy Yes

1 Sunny Yes

2 Overcast Yes

3 Overcast Yes

4 Sunny No

5 Rainy Yes

6 Sunny Yes

7 Overcast Yes

8 Rainy No

9 Sunny No

10 Sunny Yes

Faculty Name : Mrs Swapna Subject Name :ML

11 Rainy No

12 Overcast Yes

13 Overcast Yes

Frequency table for the Weather Conditions:

Weather Yes No

Overcast 5 0

Rainy 2 2

Sunny 3 2

Total 10 4

Likelihood table weather condition:

Weather No Yes

Overcast 0 5 5/14= 0.35

Rainy 2 2 4/14=0.29

Sunny 2 3 5/14=0.35

All 4/14=0.29 10/14=0.71

Applying Bayes'theorem:

Faculty Name : Mrs Swapna Subject Name :ML

P(Yes|Sunny)= P(Sunny|Yes)*P(Yes)/P(Sunny)

P(Sunny|Yes)= 3/10= 0.3

P(Sunny)= 0.35

P(Yes)=0.71

P(Yes|Sunny) = 0.3*0.71/0.35= 0.60

P(No|Sunny)= P(Sunny|No)*P(No)/P(Sunny)

P(Sunny|NO)= 2/4=0.5

P(No)= 0.29

P(Sunny)= 0.35

So P(No|Sunny)= 0.5*0.29/0.35 = 0.41

So as we can see from the above calculation that P(Yes|Sunny)>P(No|Sunny)

Hence on a Sunny day, Player can play the game.

Applications of Naïve Bayes Classifier:

o It is used for Credit Scoring.

o It is used in medical data classification.
o It is used in Text classification such as Spam filtering and Sentiment analysis.

Navie Bayers Text Classification:

pVj is the prior probablility given the data belows to that class

example : Target class : the data belongs to the class (y or no) or (like or Dislike)

P(a\vj)= Conditional probability or likelihood

Faculty Name : Mrs Swapna Subject Name :ML

8=Total number of words in 3 docid 1,2, 3=8
6=Total number of unique words count like Chinese Beijing shanghai Tokyo japan maco
3= total number of word in doc id 4

Faculty Name : Mrs Swapna Subject Name :ML

Bayesian Belief Network in artificial intelligence

Bayesian belief network is key computer technology for dealing with probabilistic events and to
solve a problem which has uncertainty. We can define a Bayesian network as:

"A Bayesian network is a probabilistic graphical model which represents a set of variables and
their conditional dependencies using a directed acyclic graph."

It is also called a Bayes network, belief network, decision network, or Bayesian model.

Real world applications are probabilistic in nature, and to represent the relationship between
multiple events, we need a Bayesian network. It can also be used in various tasks
including prediction, anomaly detection, diagnostics, automated insight, reasoning, time
series prediction, and decision making under uncertainty.

Faculty Name : Mrs Swapna Subject Name :ML

Bayesian Network can be used for building models from data and experts opinions, and it
consists of two parts:

o Directed Acyclic Graph

o Table of conditional probabilities.

The generalized form of Bayesian network that represents and solve decision problems under
uncertain knowledge is known as an Influence diagram.

A Bayesian network graph is made up of nodes and Arcs (directed links), where:

o Each node corresponds to the random variables, and a variable can

be continuous or discrete.
o Arc or directed arrows represent the causal relationship or conditional probabilities
between random variables. These directed links or arrows connect the pair of nodes in the

Faculty Name : Mrs Swapna Subject Name :ML

 graph.
These links represent that one node directly influence the other node, and if there is no
directed link that means that nodes are independent with each other
o In the above diagram, A, B, C, and D are random variables represented by
the nodes of the network graph.
o If we are considering node B, which is connected with node A by a directed
arrow, then node A is called the parent of Node B.
o Node C is independent of node A.

The Bayesian network has mainly two components:

o Causal Component
o Actual numbers

Each node in the Bayesian network has condition probability distribution P(Xi |Parent(Xi) ),
which determines the effect of the parent on that node.

Explanation of Bayesian network:

Let's understand the Bayesian network through an example by creating a directed acyclic graph:

Example: Harry installed a new burglar alarm at his home to detect burglary. The alarm reliably
responds at detecting a burglary but also responds for minor earthquakes. Harry has two
neighbors David and Sophia, who have taken a responsibility to inform Harry at work when they
hear the alarm. David always calls Harry when he hears the alarm, but sometimes he got
confused with the phone ringing and calls at that time too. On the other hand, Sophia likes to
listen to high music, so sometimes she misses to hear the alarm. Here we would like to compute
the probability of Burglary Alarm.

Problem:

Calculate the probability that alarm has sounded, but there is neither a burglary, nor an
earthquake occurred, and David and Sophia both called

Solution:

Faculty Name : Mrs Swapna Subject Name :ML

 o The Bayesian network for the above problem is given below. The network structure is
showing that burglary and earthquake is the parent node of the alarm and directly
affecting the probability of alarm's going off, but David and Sophia's calls depend on
alarm probability.
o The network is representing that our assumptions do not directly perceive the burglary
and also do not notice the minor earthquake, and they also not confer before calling.
o The conditional distributions for each node are given as conditional probabilities table

List of all events occurring in this network:

o Burglary (B)
o Earthquake(E)
o Alarm(A)
o David Calls(D)
o Sophia calls(S)

P(B= True) = 0.002, which is the probability of burglary.

Faculty Name : Mrs Swapna Subject Name :ML

P(B= False)= 0.998, which is the probability of no burglary.

P(E= True)= 0.001, which is the probability of a minor earthquake

P(E= False)= 0.999, Which is the probability that an earthquake not occurred.

We can provide the conditional probabilities as per the below tables:

Conditional probability table for Alarm A:

The Conditional probability of Alarm A depends on Burglar and earthquake:

B E P(A= True) P(A= False)

True True 0.94 0.06

True False 0.95 0.04

False True 0.31 0.69

False False 0.001 0.999

Conditional probability table for David Calls:

The Conditional probability of David that he will call depends on the probability of Alarm.

A P(D= True) P(D= False)

True 0.91 0.09

False 0.05 0.95

Conditional probability table for Sophia Calls:

The Conditional probability of Sophia that she calls is depending on its Parent Node "Alarm."

A P(S= True) P(S= False)

True 0.75 0.25

Faculty Name : Mrs Swapna Subject Name :ML

False 0.02 0.98

From the formula of joint distribution, we can write the problem statement in the form of probability distribution:

P(S, D, A, ¬B, ¬E) = P (S|A) *P (D|A) *P (¬B) *P (¬E).

= 0.75* 0.91* 0.001* 0.998*0.999

= 0.00068045.

----------------------------------------------------------------------------------------------------
Expectation-Maximization Algorithm

Expectation-Maximization algorithm can be used for the latent variables (variables that are
not directly observable and are actually inferred from the values of the other observed
variables) too in order to predict their values with the condition that the general form of
probability distribution governing those latent variables is known to us. This algorithm is
actually at the base of many unsupervised clustering algorithms in the field of machine
learning.
It was explained, proposed and given its name in a paper published in 1977 by Arthur
Dempster, Nan Laird, and Donald Rubin. It is used to find the local maximum likelihood
parameters of a statistical model in the cases where latent variables are involved and the data
is missing or incomplete.

Algorithm:
1. Given a set of incomplete data, consider a set of starting parameters.
2. Expectation step (E – step): Using the observed available data of the dataset, estimate
(guess) the values of the missing data.
3. Maximization step (M – step): Complete data generated after the expectation (E) step is
used in order to update the parameters.
4. Repeat step 2 and step 3 until convergence.

Faculty Name : Mrs Swapna Subject Name :ML

The essence of Expectation-Maximization algorithm is to use the available observed data of the dataset to
estimate the missing data and then using that data to update the values of the parameters. Let us
understand the EM algorithm in detail.
 Initially, a set of initial values of the parameters are considered. A set of incomplete observed data is
given to the system with the assumption that the observed data comes from a specific model.
 The next step is known as “Expectation” – step or E-step. In this step, we use the observed data in order
to estimate or guess the values of the missing or incomplete data. It is basically used to update the
variables.
 The next step is known as “Maximization”-step or M-step. In this step, we use the complete data
generated in the preceding “Expectation” – step in order to update the values of the parameters. It is
basically used to update the hypothesis.
 Now, in the fourth step, it is checked whether the values are converging or not, if yes, then stop otherwi se
repeat step-2 and step-3 i.e. “Expectation” – step and “Maximization” – step until the convergence occurs.

Faculty Name : Mrs Swapna Subject Name :ML

Flow chart for EM algorithm –

Usage of EM algorithm –
 It can be used to fill the missing data in a sample.
 It can be used as the basis of unsupervised learning of clusters.
 It can be used for the purpose of estimating the parameters of Hidden Markov Model (HMM).
 It can be used for discovering the values of latent variables.
Advantages of EM algorithm –
 It is always guaranteed that likelihood will increase with each iteration.
 The E-step and M-step are often pretty easy for many problems in terms of implementation.
Disadvantages of EM algorithm –
 It has slow convergence.
 It makes convergence to the local optima only.
.

Probably Approximately Correct (PAC) framework, we identify classes of hypotheses that can and
cannot be learned from a polynomial number of training examples and we define a natural measure of
complexity for hypothesis spaces that allows bounding the number of training examples required for
learning. Within the mistake bound framework, we examine the number of training errors that will be
made by a learner before it determines the correct hypothesis

Faculty Name : Mrs Swapna Subject Name :ML

 we will be chiefly concerned with questions such as how many training examples are sufficient to
successfully learn the target function, and how many mistakes will the learner make before succeeding.
As we shall see, it is possible to set quantitative bounds on these measures, depending on attributes of
the learning problem such as:

1) the size or complexity of the hypothesis space considered by the learner

2) the accuracy to which the target concept must be approximated
3) the probability that the learner will output a successful hypothesis
4) the manner in which training examples are presented to the learner

Goal of PAC is to answer questions such as:

1.Sample complexity. How many training examples are needed for a learner to converge (with
high probability) to a successful hypothesis?

2. Computational complexity. How much computational effort is needed for a learner to

converge (with high probability) to a successful hypothesis?

3. Mistake bound. How many training examples will the learner misclassify before converging to
a successful hypothesis?

Error of a Hypothesis

Acccuracy is based on Error Levels in the model

This is made clear by distinguishing between the true error of a model and the estimated or
sample error.

 Sample Error. Estimate of error calculated on a sample data.

 The sample error (errors(h)) of hypothesis h with respect to target function f anddata
sample S is

Where n is the number of examples in S, and the quantity δ(f(x), h(x)) is 1 if error is
identified

Faculty Name : Mrs Swapna Subject Name :ML

 if f (x) ≠ h(x), and 0 no error identified.
Suppose the data sample S contains n = 40 examples and that hypothesis h
commits r =12 errors misclassify or mismatch over this data.
 The sample error is errors(h) = r/n = 12/40 = 0.30

 True Error: Estimation of Error over entire distribution

 The true error (errorD(h)) of hypothesis h with respect to target function f and
distribution D, is the probability that h will misclassify an instance drawn at random
accordingto D.

refer to unit 2 calculation of true error through confidence intervals

Confidence Intervals for Discrete-Valued Hypotheses

Suppose we wish to estimate the true error for some discrete valued hypothesis h, based on its
observed sample error over a sample S, where
 The sample S contains n examples drawn independent of one another, and independent of h,
according to the probability distribution D
 n ≥ 30
 Hypothesis h commits r errors over these n examples (i.e., errors (h) = r/n).

Under these conditions, statistical theory allows to make the following assertions:
2. Given no other information, the most probable value of errorD (h) is errors(h)
3. With approximately 95% probability, the true error errorD (h) lies in the interval

Faculty Name : Mrs Swapna Subject Name :ML

 Example:

Suppose the data sample S contains n = 40 examples and that hypothesis h commits r = 12
errors over this data.
 The sample error is errors(h) = r/n = 12/40 = 0.30
 Given no other information, true error is errorD (h) = errors(h), i.e., errorD (h) =0.30
 With the 95% confidence interval estimate for errorD (h).

= 0.30 ± (1.96 * 0.07) = 0.30 ± 0.14

True Error vs Sample Error

True Error Sample Error

The true error represents the

probability that a random sample
from the population is Sample Error represents the fraction of the sample which
misclassified. is misclassified.

True error is used to estimate the

error of the population. Sample Error is used to estimate the errors of the sample.

True error is difficult to calculate.

It is estimated by the confidence
interval range on the basis of Sample Error is easy to calculate. You just have to
Sample error. calculate the fraction of the sample that is misclassified.

The true error can be caused by Sampling error can be of type population-specific error
poor data collection methods, (wrong people to survey), selection error, sample-frame

Faculty Name : Mrs Swapna Subject Name :ML

True Error Sample Error

selection bias, or non-response error (wrong frame window selected for sample), and
bias. non-response error (when respondent failed to respond).

PROBABLY LEARNING AN APPROXIMATELY CORRECT HYPOTHESIS

We begin by specifying the problem setting that defines the PAC learning model, then consider the
questions of how many training examples and how much computation are required in order to learn
various classes of target functions within this PAC model.

PAC-learnability is largely determined by the number of training examples required by the learner. The
growth in the number of required training examples with problem size, called the sample complexity of
the learning problem

a general bound on the sample complexity for a very broad class of learners, called consistent learners.
A learner is consistent if it outputs hypotheses that perfectly fit the training data.

The learner L considers some set H of possible hypotheses when attempting to learn the target concept.
For example, H might be the set of all hypotheses describable by conjunctions of the attributes. After
observing a sequence of training examples of the target concept c, L must output some hypothesis h
from H, which is its estimate of c.

To be fair, we evaluate the success of L by the performance of h over new instances drawn randomly
from X according to D(Traning data), the same probability distribution used to generate the training data

Error of a Hypothesis

Figure 7.1 shows this definition of error in graphical form. The concepts c and h are depicted by the sets
of instances within X that they label as positive. The error of h with respect to c is the probability that a
randomly drawn instance will fall into the region where h and c disagree

Faculty Name : Mrs Swapna Subject Name :ML

PAC Learnability

Our aim is to characterize classes of target concepts that can be reliably learned from a reasonable
number of randomly drawn training examples and a reasonable amount of computation

First, unless we provide training examples corresponding to every possible instance in X (an unrealistic
assumption), there may be multiple hypotheses consistent with the provided training examples, and the
learner cannot be certain to pick the one corresponding to the target concept. Second, given that the
training examples are drawn randomly, there will always be some nonzero probability that the training
examples encountered by the learner will be misleading

To accommodate these two difficulties, we weaken our demands on the learner in two ways. First, we
will not require that the learner output a zero error hypothesis-we will require only that its error be
bounded by some constant, c, that can be made arbitrarily small.

Faculty Name : Mrs Swapna Subject Name :ML

SAMPLE COMPLEXITY FOR FINITE HYPOTHESIS SPACES

The growth in the number of required training examples with problem size, called the sample
complexity of the learning problem

we present a general bound on the sample complexity for a very broad class of learners, called
consistent learners. A learner is consistent if it outputs hypotheses that perfectly fit the training
data

Find s Algorithm –biased Hypothesis

Candidate Algorithm –Unbiased hypothesis Restricted hypothesis

DecisonTree – preference Bias

The significance of the version space here is that every consistent learner outputs a hypothesis
belonging to the version space, regardless of the instance space X, hypothesis space H, or
training data D. The reason is simply that by definition the version space VSH,D contains every

Faculty Name : Mrs Swapna Subject Name :ML

consistent hypothesis in H. Therefore, to bound the number of examples needed by any
consistent learner

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
SAMPLE COMPLEXITY FOR INFINITE HYPOTHESIS SPACES

Sample Complexity Results for Infinite Hypothesis Spaces can be explained with concept of
shattering coefficient

The Shattering Coefficient Let C be a concept class over an instance space X, i.e. a set of
functions functions from X to {0, 1} (where both C and X may be infinite).

Shattering a Set of Instances

Faculty Name : Mrs Swapna Subject Name :ML

The Vapnik-Chervonenkis Dimension

Faculty Name : Mrs Swapna Subject Name :ML

The VC dimension quantifies the complexity of a hypothesis space, e.g. the models that could be
fit given a representation and learning

The ability to shatter a set of instances is closely related to the inductive bias of a hypothesis
space. Algorithm

Shatter or a shattered set in the case of a dataset, means points in the feature space can be
selected or separated from each other using hypotheses in the space such that the labels of
examples in the separate groups are correct

THE MISTAKE BOUND MODEL OF LEARNING

the mistake bound model of learning, in which the learner is evaluated by the total number of mistakes
it makes before it converges to the correct hypothesis. As in the PAC setting, we assume the learner
receives a sequence of training examples.

However, here we demand that upon receiving each example x, the learner must predict the target
value c(x), before it is shown the correct target value by the trainer. The question considered is "How
many mistakes will the learner make in its predictions before it learns the target concept?' This question
is significant in practical settings where learning must be done while the system is in actual use, rather
than during some off-line training stage.

For example, if the system is to learn to predict which credit card purchases should be approved and
which are fraudulent, based on data collected during use, then we are interested in minimizing the total
number of mistakes it will make before converging to the correct target function. Here the total number
of mistakes can be even more important than the total number of training examples.

This mistake bound learning problem may be studied in various specific settings. For example, we might
count the number of mistakes made before PAC learning the target concept. In the examples below, we

Faculty Name : Mrs Swapna Subject Name :ML

consider instead the number of mistakes made before learning the target concept exactly. Learning the
target concept exactly means converging to a hypothesis such that (Vx)h(x) = c(x).

Mistake Bound for the FIND-S Algorithm

To illustrate, consider again the hypothesis space H consisting of conjunctions of up to n boolean literals
l1 , l2…ln, and their negations Recall the FIND-S algorithm , which incrementally computes the maximally
specific hypothesis consistent with the training examples. A straightforward implementation of FIND-S
for the hypothesis space H is as follow

FIND-S converges in the limit to a hypothesis that makes no errors, provided C, H and provided the
training data is noise-free. FIND-S begins with the most specific hypothesis (which classifies every
instance a negative example), then incrementally generalizes this hypothesis as needed to cover
observed positive training examples. For the hypothesis representation used here, this generalization
step consists of deleting unsatisfied literals.

Therefore, to calculate the number of mistakes it will make, we need only count the number of mistakes
it will make misclassifying truly positive examples as negative.

Therefore, to calculate the number of mistakes it will make, we need only count the number of mistakes
it will make misclassifying truly positive examples as negative.

How many such mistakes can occur before FIND-S learns c exactly? Consider the first positive example
encountered by FIND-S. The learner will certainly make a mistake classifying this example, because its
initial hypothesis labels every instance negative. However, the result will be that 1/ 2n terms in its initial
hypothesis will be eliminated, leaving only n terms. For each subsequent positive example that is
mistakenly classified by the current hypothesis, at least one more of the remaining n terms must be
eliminated from the hypothesis.

Faculty Name : Mrs Swapna Subject Name :ML

Therefore, the total number of mistakes can be at most n + 1. This number of mistakes will be required
in the worst case, corresponding to learning the most general possible target concept .

---------------------------------------------------------------------------------------------------------------------------------------

INSTANCE BASED LEARNING

 Instance-based learning methods such as nearest neighbor and locally weighted

regression are conceptually straightforward approaches to approximating real-valued or
discrete-valued target functions.
 Learning in these algorithms consists of simply storing the presented training data.
When a new query instance is encountered, a set of similar related instances is retrieved
from memory and used to classify the new query instance
 Instance-based approaches can construct a different approximation to the target function
for each distinct query instance that must be classified

The Machine Learning systems which are categorized as instance-based learning are the
systems that learn the training examples by heart and then generalizes to new instances based
on some similarity measure. It is called instance-based because it builds the hypotheses from
the training instances. It is also known as memory-based learning or lazy-learning. The time
complexity of this algorithm depends upon the size of training data.

k- NEAREST NEIGHBOR LEARNING

o K-Nearest Neighbour is one of the simplest Machine Learning algorithms based on

Supervised Learning technique.
o K-NN algorithm assumes the similarity between the new case/data and available cases
and put the new case into the category that is most similar to the available categories.
o K-NN algorithm stores all the available data and classifies a new data point based on the
similarity. This means when new data appears then it can be easily classified into a well
suite category by using K- NN algorithm.
o K-NN algorithm can be used for Regression as well as for Classification but mostly it is
used for the Classification problems.

 The most basic instance-based method is the K- Nearest Neighbor Learning. This
algorithm assumes all instances correspond to points in the n-dimensional space Rn.

Faculty Name : Mrs Swapna Subject Name :ML

  The nearest neighbors of an instance are defined in terms of the standard Euclidean
distance.
 Let an arbitrary instance x be described by the feature vector
((a1(x), a2(x), ………, an(x))
Where, ar(x) denotes the value of the rth attribute of instance x.

 Then the distance between two instances xi and xj is defined to be d(xi ,

xj ) Where,

 In nearest-neighbor learning the target function may be either discrete-valued

or real- valued.

Let us first consider learning discrete-valued target functions of

the form Where, V is the finite set {v1, . . . vs }

 The value �(xq) returned by this algorithm as its estimate of f(x q) is just the
most common value of f among the k training examples nearest to xq.
 If k = 1, then the 1- Nearest Neighbor algorithm assigns to �(xq) the value f(xi).
Where xi is the training instance nearest to xq.
 For larger values of k, the algorithm assigns the most common value among the k
nearest training examples.

 Below figure illustrates the operation of the k-Nearest Neighbor algorithm for the
case where the instances are points in a two-dimensional space and where the
target function is Boolean valued.

Faculty Name : Mrs Swapna Subject Name :ML

  The positive and negative training examples are shown by “+” and “-”
respectively. A query point xq is shown as well.
 The 1-Nearest Neighbor algorithm classifies xq as a positive example in this
figure, whereas the 5-Nearest Neighbor algorithm classifies it as a negative
example.

 Below figure shows the shape of this decision surface induced by 1- Nearest
Neighbor over the entire instance space. The decision surface is a combination of
convex polyhedra surrounding each of the training examples.

 For every training example, the polyhedron indicates the set of query points
whose classification will be completely determined by that training example.
Query points outside the polyhedron are closer to some other training
example. This kind of diagram is often called the Voronoi diagram of the set of
training example

Example of K Nearest

Faculty Name : Mrs Swapna Subject Name :ML

Advantages of KNN Algorithm:
o It is simple to implement.
o It is robust to the noisy training data
o It can be more effective if the training data is large.

Disadvantages of KNN Algorithm:

o Always needs to determine the value of K which may be complex some time.
o The computation cost is high because of calculating the distance between the data points for all the training
samples.

Faculty Name : Mrs Swapna Subject Name :ML

Example of knn Problem

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
 The k- Nearest Neighbor algorithm for approximation a discrete-valued target function
is
given below:

Faculty Name : Mrs Swapna Subject Name :ML

 The K- Nearest Neighbor algorithm for approximation a real-valued target
function is given below

Faculty Name : Mrs Swapna Subject Name :ML

 Distance-Weighted Nearest Neighbor Algorithm for approximation a discrete-valued
target functions

Faculty Name : Mrs Swapna Subject Name :ML

 o

Faculty Name : Mrs Swapna Subject Name :ML

 Distance-Weighted Nearest Neighbor Algorithm for approximation a Real-valued target
functions

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
Regression Analysis in Machine learning
Regression analysis is a statistical method to model the relationship between a dependent (target) and independent
(predictor) variables with one or more independent variables. More specifically, Regression analysis helps us to
understand how the value of the dependent variable is changing corresponding to an independent variable when
other independent variables are held fixed. It predicts continuous/real values such as temperature, age, salary,
price, etc.

We can understand the concept of regression analysis using the below example:

Example: Suppose there is a marketing company A, who does various advertisement every year and get sales on that.
The below list shows the advertisement made by the company in the last 5 years and the corresponding sales:

Now, the company wants to do the advertisement of $200 in the year 2019 and wants to know the prediction
about the sales for this year. So to solve such type of prediction problems in machine learning, we need regression
analysis.

Linear Regression in Machine Learning

Linear regression is one of the easiest and most popular Machine Learning algorithms. It is a statistical method that is
used for predictive analysis. Linear regression makes predictions for continuous/real or numeric variables such
as sales, salary, age, product price, etc.

Faculty Name : Mrs Swapna Subject Name :ML

Linear regression algorithm shows a linear relationship between a dependent (y) and one or more independent (x)
variables, hence called as linear regression. Since linear regression shows the linear relationship, which means it finds
how the value of the dependent variable is changing according to the value of the independent variable.

The linear regression model provides a sloped straight line representing the relationship between the variables.
Consider the below image:

Mathematically, we can represent a linear regression as:

y= a0+a1x+ ε
Y= Dependent Variable (Target Variable)
X= Independent Variable (predictor Variable)
a0= intercept of the line (Gives an additional degree of freedom)
a1 = Linear regression coefficient (scale factor to each input value).
ε = random error

The values for x and y variables are training datasets for Linear Regression model representation.

2.6M
Obi-Wan Is Coming to ‘
for Linear Regression, we use the Mean Squared Error (MSE) cost function, which is the average of squared error
occurred between the predicted values and actual values. It can be written as:

For the above linear equation, MSE can be calculated as:

Faculty Name : Mrs Swapna Subject Name :ML

Where,

N=Total number of observation

Yi = Actual value
(a1xi+a0)= Predicted value.

1.LOCALLY WEIGHTED REGRESSION

 The phrase "locally weighted regression" is called local because the function is
approximated based only on data near the query point, weighted because the
contribution of each training example is weighted by its distance from the
query point, and regression because this is the term used widely in the
statistical learning community for the problem of approximating real-valued
functions.

 Given a new query instance xq, the general approach in locally weighted
regression is to construct an approximation 𝑓 that fits the training examples in
the neighborhood surrounding xq. This approximation is then used to calculate
the value 𝑓(xq), which is output as the estimated target value for the query
instance.
 Consider locally weighted regression in which the target function f is

approximated near xq using a linear function of the form

Where, ai(x) denotes the value of the ith attribute of the instance x

 Derived methods are used to choose weights that minimize the squared error
summed over the set D of training examples using gradient descent

Which led us to the gradient descent training rule

Faculty Name : Mrs Swapna Subject Name :ML

 Where, η is a constant learning rate

Need to modify this procedure to derive a local approximation rather than a global one.
The simple way is to redefine

2.Radial basis functions

A radial basis function network is a type of supervised artificial neural network that uses
supervised machine learning (ML) to function as a nonlinear classifier
A radial basis function network is also known as a radial basis network.

The radial basis function network uses radial basis functions as its activation functions. Like
other kinds of neural networks, radial basis function networks have input layers, hidden layers
and output layers. However, radial basis function networks often also include a nonlinear
activation function of some kind. Output weights can be trained using gradient descent.

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
3.Case-based reasoning (CBR)

Case-based reasoning (CBR) is an experience-based approach to solving new problems by adapting

previously successful solutions to similar problems. Addressing memory, learning, planning and problem
solving, CBR provides a foundation for a new technology of intelligent computer systems that can solve
problems and adapt to new situations. In CBR, the “intelligent” reuse of knowledge from already-solved
problems
Four step process for CBR

In general, the case-based reasoning process entails:

1. Retrieve- Gathering from memory an experience closest to the current problem.

2. Reuse- Suggesting a solution based on the experience and adapting it to meet the demands of the

new situation.

3. Revise- Evaluating the use of the solution in the new context.

4. Retain- Storing this new problem-solving method in the memory system.

Faculty Name : Mrs Swapna Subject Name :ML

Advantages and disadvantages of CBR

On the plus side, remembering past experiences helps learners avoid repeating previous mistakes, and
the reasoned can discern what features of a problem are significant and focus on them.

On the negative side, critics claim that the main premise of CBR is based on anecdotal evidence and that
adapting the elements of one case to another may be complex and potentially lead to inaccuracies

REMARKS ON LAZY AND EAGER LEARNING

Faculty Name : Mrs Swapna Subject Name :ML

Lazy learner:

1. Just store Data set without learning from it

2. Start classifying data when it receive Test data
3. So it takes less time learning and more time classifying data

Eager learner:

1. When it receive data set it starts classifying (learning)

2. Then it does not wait for test data to learn
3. So it takes long time learning and less time classifying data

Lazy : K - Nearest Neighbour, Case - Based Reasoning

Eager : Decision Tree, Naive Bayes, Artificial Neural Networks

Faculty Name : Mrs Swapna Subject Name :ML

Unit 4
Genetic Algorithms

Genetic Algorithms(GAs) are adaptive heuristic search algorithms that belong to the larger part of
evolutionary algorithms. Genetic algorithms are based on the ideas of natural selection and genetics.
These are intelligent exploitation of random search provided with historical data to direct the search
into the region of better performance in solution space. They are commonly used to generate high-
quality solutions for optimization problems and search problems.
Genetic algorithms simulate the process of natural selection which means those species who can
adapt to changes in their environment are able to survive and reproduce and go to next generation. In
simple words, they simulate “survival of the fittest” among individual of consecutive generation for
solving a problem. Each generation consist of a population of individuals and each individual
represents a point in search space and possible solution. Each individual is represented as a string of
character/integer/float/bits. This string is analogous to the Chromosome.

Five phases are considered in a genetic algorithm.

• Initial population
• Fitness function
• Selection
• Crossover
• Mutation
• The process begins with a set of individuals which is called a Population. Each individual is a
solution to the problem you want to solve.
• An individual is characterized by a set of parameters (variables) known as Genes. Genes are
joined into a string to form a Chromosome (solution).
• In a genetic algorithm, the set of genes of an individual is represented using a string, in terms of
an alphabet. Usually, binary values are used (string of 1s and 0s). We say that we encode the
genes in a chromosome
• The fitness function determines how fit an individual is (the ability of an individual to compete
with other individuals). It gives a fitness score to each individual. The probability that an
individual will be selected for reproduction is based on its fitness score.

Selection
Faculty Name : Mrs Swapna Subject Name :ML
 • The idea of selection phase is to select the fittest individuals and let them pass their genes to
the next generation.
• Two pairs of individuals (parents) are selected based on their fitness scores. Individuals with
high fitness have more chance to be selected for reproduction.

Crossover

• Crossover is the most significant phase in a genetic algorithm. For each pair of parents to be
mated, a crossover point is chosen at random from within the genes.

Mutation

• In certain new offspring formed, some of their genes can be subjected to a mutation with a low
random probability. This implies that some of the bits in the bit string can be flipped.

Faculty Name : Mrs Swapna Subject Name :ML

Offspring are created by exchanging the genes of parents among themselves until the crossover point is
reached.

Faculty Name : Mrs Swapna Subject Name :ML

The new offspring are added to the population.

In certain new offspring formed, some of their genes can be subjected to a mutation with a low random
probability. This implies that some of the bits in the bit string can be flipped.

Termination

The algorithm terminates if the population has converged (does not produce offspring which are
significantly different from the previous generation). Then it is said that the genetic algorithm has
provided a set of solutions to our problem.
Faculty Name : Mrs Swapna Subject Name :ML
 Application areas

Genetic algorithms are applied in the following fields:

 Transport: Genetic algorithms are used in the traveling salesman problem to develop transport
plans that reduce the cost of travel and the time taken. They are also used to develop an efficient
way of delivering products.
 DNA Analysis: They are used in DNA analysis to establish the DNA structure using
spectrometric information.
 Multimodal Optimization: They are used to provide multiple optimum solutions in multimodal
optimization problems.
 Aircraft Design: They are used to develop parametric aircraft designs. The parameters of the
aircraft are modified and upgraded to provide better designs.
 Economics: They are used in economics to describe various models such as the game theory,
asset pricing, and schedule optimization.

Advantages of Genetic Algorithms

 Parallelism
 A larger set of solution space
 Requires less information
 Provides multiple optimal solutions
 Probabilistic in nature
 Genetic representations using chromosomes

Disadvantages of Genetic Algorithms

 Computational complexity

Hypothesis Search space

 The population of individuals are maintained within search space. Each individual represents a
solution in search space for given problem. Each individual is coded as a finite length vector

Faculty Name : Mrs Swapna Subject Name :ML

(analogous to chromosome) of components. These variable components are analogous to Genes. Thus
a chromosome (individual) is composed of several genes (variable components).

Genetic Programming Example

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
Models of evolution and learning proposed by 2 authors

Lamarckian Evolution Theory:

The Lamarckian theory states the characteristic individual acquire during their lifetime pass them to
their children. This theory is named after French biologist Jean Baptiste Lamarck. According to
Lamarck‟s theory, learning is an important part of the evolution of species(or for our purpose in the
Evolutionary algorithm). This theory is discredited in a biological context but can be used in genetic
algorithms in machine learning.

Baldwin Effect:
Baldwin proposed that individual learning can explain evolutionary phenomena that appear to require
Lamarckian inheritance of acquired characteristics. The ability of individuals to learn can guide the
evolutionary process. In effect, learning smooths the fitness landscape, thus facilitating evolution.

Faculty Name : Mrs Swapna Subject Name :ML

 Baldwin Effect is first demonstrated by Hinton and Nolan in the context of machine learning in 1987.
They take simple Neural Networks (NNs). In one experiment they take NNs of fixed weights while
other NNs set to trainable. They concluded that:
 When there is no individual learning, the population(collection of NNs) failed to improve over
time.
 When learning is applied in early stages, the population contains many individuals with many
trainable weights, but in later stages, it achieved high fitness with the number of trainable weights
decreases in individuals

 Parallelizing Genetic Algorithms

 parallel genetic algorithm is such an algorithm that uses multiple genetic algorithms to solve a
single task . All these algorithms try to solve the same task and after they‟ve completed their job,
the best individual of every algorithm is selected, then the best of them is selected, and this is the
solution to a problem. This is one of the most popular approach to parallel genetic algorithms,
even though there are others. This approach is often called „island model‟ because populations are
isolated from each other, like real-life creature populations may be isolated living on different
islands. Image 1 illustrates that.

These genetic algorithms do not depend on each other, as a result, they can run in parallel, taking
advantage of a multicore CPU. Each algorithm has its own set of individual, as a result these individuals
may differ from individuals of another algorithm, because they have different mutation/crossover history.

Faculty Name : Mrs Swapna Subject Name :ML

Let‟s illustrate this with an example. Say we have three independent genetic algorithms and we want to
crossover them in pairs. We take the first algorithm and randomly select the second element of a pair. So,
we create as much pairs as many algorithms we have, in each pair the first element is chosen sequentially,
and the second one is random. Then we perform crossover on populations of these two algorithms taking
individuals from both of them. We pick individuals for crossover in such a way, that individuals from
different algorithms are crossed over together. We use crossover mechanisms that are used by the first
algorithm of a pair and this algorithm receives all of the individuals that were created as a result of a
crossover; the second algorithm of a pair is simply a donor that provides its individuals. Therefore, each
algorithm receives new individuals when it is the first element of a pair. If a crossover algorithm requires
more than two individuals, additional individuals may be taken from any of the algorithms of a pair, it is
only recommended that no algorithm dominate here

A parallel genetic algorithm may take a little more time than a non-parallel one, that is because is uses
several computation threads which, in turn, cause the Operation System to perform context switching
more frequently

Distributed genetic algorithm

Distributed genetic algorithm is actually a parallel genetic algorithm that has its independent algorithms
running on separate machines. Moreover, in this case each of these algorithms may be in turn a parallel
genetic algorithm! Distributed genetic algorithm also implements the „island model‟ and each „island‟ is
even more isolated from others. If each machine runs a parallel genetic algorithm we may call this as
„archipelago model‟, because we have groups of islands. It actually does not matter what a single genetic
algorithm is, because distributed genetic algorithm is about having multiple machines running
independent genetic algorithms in order to solve the same task. Image 2 illustrates this.

Faculty Name : Mrs Swapna Subject Name :ML

Distributed genetic algorithm may also help when we have to create many individuals in order to observe
the entire domain, but it is not possible to store all of them in memory of a single machine.

When we were discussing parallel genetic algorithm we introduced the „crossover between algorithms‟
term. Distributed genetic algorithm enables us to perform crossover between separate machines

In case of distributed genetic algorithm, we have a kind of „master mind‟ that controls the overall
progress and coordinates these machines. It also controls crossover between machines, selecting how
machines will be paired together to perform crossover. In general, process is the same as in case of
parallel genetic algorithm, except that individuals are moved over the network from one machine to
another.

Sequential Covering Algorithm

Sequential Covering is a popular algorithm based on Rule-Based Classification used for learning a
disjunctive set of rules. The basic idea here is to learn one rule, remove the data that it covers, then
repeat the same process. In this process, In this way, it covers all the rules involved with it in a
sequential manner during the training phase.
The Sequential Learning algorithm takes care of to some extent, the low coverage problem in the
Learn-One-Rule algorithm covering all the rules in a sequential manner.

Working on the Algorithm:

Faculty Name : Mrs Swapna Subject Name :ML
The algorithm involves a set of „ordered rules‟ or „list of decisions‟ to be made.

Step 1 – create an empty decision list, ‘R’.

Step 2 – ‘Learn-One-Rule’ Algorithm
Step 2.a – if all training examples ∈ class ‘y’, then it’s classified as positive example.
Step 2.b – else if all training examples ∉ class ‘y’, then it’s classified as negative example.

Step 3 – The rule becomes ‘desirable’ when it covers a majority of the positive examples.
Step 4 – When this rule is obtained, delete all the training data associated with that rule.
(i.e. when the rule is applied to the dataset, it covers most of the training data, and has to be removed)
Step 5 – The new rule is added to the bottom of decision list, ‘R’. (Fig.3)

Below, is a visual representation describing the working of the algorithm.

Faculty Name : Mrs Swapna Subject Name :ML

 fig 4: Visual Representation of working of the algorithm
 Let us understand step by step how the algorithm is working in the example shown in Fig.4.
 First, we created an empty decision list. During Step 1, we see that there are three sets of positive
examples present in the dataset. So, as per the algorithm, we consider the one with maximum no of
positive example. (6, as shown in Step 1 of Fig 4)
 Once we cover these 6 positive examples, we get our first rule R 1, which is then pushed into the
decision list and those positive examples are removed from the dataset. (as shown in Step 2 of Fig
4)
 Now, we take the next majority of positive examples (5, as shown in Step 2 of Fig 4) and follow the
same process until we get rule R 2. (Same for R3)
 In the end, we obtain our final decision list with all the desirable rules.

Propositional logic:

• Propositional logic (PL) is the simplest form of logic where all the statements are made by
propositions. A proposition is a declarative statement which is either true or false. It is a
Propositional logic is also called Boolean logic as it works on 0 and 1.
• In propositional logic, we use symbolic variables to represent the logic, and we can use any
symbol for a representing a proposition, such A, B, C, P, Q, R, etc.
• Propositions can be either true or false, but it cannot be both.
• Propositional logic consists of an object, relations or function, and logical connectives.
• These connectives are also called logical operators.
• The propositions and connectives are the basic elements of the propositional logic.
Faculty Name : Mrs Swapna Subject Name :ML
 • Connectives can be said as a logical operator which connects two sentences.
• A proposition formula which is always true is called tautology, and it is also called a valid
sentence.
• A proposition formula which is always false is called Contradiction.
• A proposition formula which has both true and false values is called
• Statements which are questions, commands, or opinions are not propositions such as "Where is
Rohini", "How are you", "What is your name", are not propositions.
• technique of knowledge representation in logical and mathematical form.
• Syntax of propositional logic:
• The syntax of propositional logic defines the allowable sentences for the knowledge
representation. There are two types of Propositions:
• Atomic Propositions
• Compound propositions
• Atomic Proposition: Atomic propositions are the simple propositions. It consists of a single
proposition symbol. These are the sentences which must be either true or false.
• Example:
• a) 2+2 is 4, it is an atomic proposition as it is a true fact.
• b) "The Sun is cold" is also a proposition as it is a false fact.
• Compound proposition: Compound propositions are constructed by combining simpler or
atomic propositions, using parenthesis and logical connectives.
• Example:
• a) "It is raining today, and street is wet."
• b) "Ankit is a doctor, and his clinic is in Mumbai."
• Limitations of Propositional logic:
• We cannot represent relations like ALL, some, or none with propositional logic. Example:
• All the girls are intelligent.
• Some apples are sweet.
• Propositional logic has limited expressive power.
• In propositional logic, we cannot describe statements in terms of their properties or logical
relationships.

First-Order Logic

Faculty Name : Mrs Swapna Subject Name :ML

 • In the topic of Propositional logic, we have seen that how to represent statements using
propositional logic. But unfortunately, in propositional logic, we can only represent the facts,
which are either true or false. PL is not sufficient to represent the complex sentences or natural
language statements. The propositional logic has very limited expressive power. Consider the
following sentence, which we cannot represent using PL logic.
• "Some humans are intelligent", or
• "Sachin likes cricket."
• To represent the above statements, PL logic is not sufficient, so we required some more powerful
logic, such as first-order logic.
• First-Order logic:
• First-order logic is another way of knowledge representation in artificial intelligence. It is an
extension to propositional logic.
• FOL is sufficiently expressive to represent the natural language statements in a concise way.
• First-order logic is also known as Predicate logic or First-order predicate logic. First-order
logic is a powerful language that develops information about the objects in a more easy way and
can also express the relationship between those objects.
• First-order logic (like natural language) does not only assume that the world contains facts like
propositional logic but also assumes the following things in the world:
• Objects: A, B, people, numbers, colors, wars, theories, squares, pits, wumpus, ......
• Relations: It can be unary relation such as: red, round, is adjacent, or n-any relation such
as: the sister of, brother of, has color, comes between
• Function: Father of, best friend, third inning of, end of, ......
• As a natural language, first-order logic also has two main parts:
• Syntax

First-order logic statements can be divided into two parts:

• Subject: Subject is the main part of the statement.

• Predicate: A predicate can be defined as a relation, which binds two atoms together in a
statement.
• Consider the statement: "x is an integer.", it consists of two parts, the first part x is the subject of
the statement and second part "is an integer," is known as a predicate.

Faculty Name : Mrs Swapna Subject Name :ML

 Quantifiers in First-order logic:

A quantifier is a language element which generates quantification, and quantification specifies the
quantity of specimen in the universe of discourse.

These are the symbols that permit to determine or identify the range and scope of the variable in the
logical expression. There are two types of quantifier:

Universal Quantifier, (for all, everyone, everything)

Existential quantifier, (for some, at least one).

Universal Quantifier:

• Universal quantifier is a symbol of logical representation, which specifies that the statement within its
range is true for everything or every instance of a particular thing.

The Universal quantifier is represented by a symbol ∀, which resembles an inverted A.

• f x is a variable, then ∀x is read as:

• For all x
• For each x
• For every x.
• Example:
• All man drink coffee.
• ∀x man(x) → drink (x, coffee).

Faculty Name : Mrs Swapna Subject Name :ML

it will be read as: There are all x where x is a man who drink coffee.

Existential Quantifier:

Existential quantifiers are the type of quantifiers, which express that the statement within its scope is true
for at least one instance of something.

It is denoted by the logical operator ∃, which resembles as inverted E. When it is used with a predicate
variable then it is called as an existential quantifier.

If x is a variable, then existential quantifier will be ∃x or ∃(x). And it will be read as:

There exists a 'x.'

For some 'x.'

For at least one 'x.'

Example:

Faculty Name : Mrs Swapna Subject Name :ML

 Some boys are intelligent

∃x: boys(x) ∧ intelligent(x)

It will be read as: There are some x where x is a boy who is intelligent.

• Some Examples of FOL using quantifier:

• 1. All birds fly.
In this question the predicate is "fly(bird)."
And since there are all birds who fly so it will be represented as follows.
∀x bird(x) →fly(x).
• 2. Every man respects his parent.
In this question, the predicate is "respect(x, y)," where x=man, and y= parent.
Since there is every man so will use ∀, and it will be represented as follows:
∀x man(x) → respects (x, parent).
• 3. Some boys play cricket.
In this question, the predicate is "play(x, y)," where x= boys, and y= game. Since there are some boys so
we will use ∃, and it will be represented as:
∃x boys(x) → play(x, cricket).
Faculty Name : Mrs Swapna Subject Name :ML
 • 4. Not all students like both Mathematics and Science.
In this question, the predicate is "like(x, y)," where x= student, and y= subject.
Since there are not all students, so we will use ∀ with negation, so following representation for
this:
¬∀ (x) [ student(x) → like(x, Mathematics) ∧ like(x, Science)].
• First Order Inductive Learner (FOIL)
• In machine learning, first-order inductive learner (FOIL) is a rule-based learning algorithm. It
is a natural extension of SEQUENTIAL-COVERING and LEARN-ONE-RULE algorithms.
• Inductive Learning:
• Inductive learning analyzing and understanding the evidence and then using it to determine the
outcome. It is based on Inductive Logic.

Faculty Name : Mrs Swapna Subject Name :ML

 Induction as inverted deduction:

A different approach to inductive logic programming is based on the simple observation that
induction is just the inverse of deduction.

INVERTING RESOLUTION :A general method for automated deduction is the resolution rule
introduced by Robinson (1965). The resolution rule is a sound and complete rule for deductive
inference in first-order logic.

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
Faculty Name : Mrs Swapna Subject Name :ML
Faculty Name : Mrs Swapna Subject Name :ML
Faculty Name : Mrs Swapna Subject Name :ML
REINFORCEMENT LEARNING

Reinforcement learning addresses the question of how an autonomous agent that senses and acts in
its environment can learn to choose optimal actions to achieve its goals. In general, a reinforcement
learning agent is able to perceive and interpret its environment, take actions and learn through trial
and error.

Faculty Name : Mrs Swapna Subject Name :ML

INTRODUCTION

 Consider building a learning robot. The robot, or agent, has a set of sensors to observe the
state of its environment, and a set of actions it can perform to alter this state.
 Its task is to learn a control strategy, or policy, for choosing actions that achieve its goals.
 The goals of the agent can be defined by a reward function that assigns a numerical value to
each distinct action the agent may take from each distinct state.
 This reward function may be built into the robot, or known only to an external teacher who
provides the reward value for each action performed by the robot.
 The task of the robot is to perform sequences of actions, observe their consequences, and
learn a control policy.
 The control policy is one that, from any initial state, chooses actions that maximize the reward
accumulated over time by the agent.

Example:

 A mobile robot may have sensors such as a camera and sonars, and actions such as "move
forward" and "turn."
 The robot may have a goal of docking onto its battery charger whenever its battery level is low.
 The goal of docking to the battery charger can be captured by assigning a positive reward
(Eg., +100) to state-action transitions that immediately result in a connection to the charger

Faculty Name : Mrs Swapna Subject Name :ML

the difference table between RL and Supervised learning is given below:

Reinforcement Learning Supervised Learning(ACTIVE LEARNING)

RL works by interacting with the environment. Supervised learning works on the existing dataset.

The RL algorithm works like the human brain works Supervised Learning works as when a human learns th
when making some decisions. the supervision of a guide.

There is no labeled dataset is present The labeled dataset is present.

No previous training is provided to the learning agent. Training is provided to the algorithm so that it can pre
output.

RL helps to take decisions sequentially. In Supervised learning, decisions are made when i
given.

Types of Reinforcement: There are two types of Reinforcement:

1. Positive –
Positive Reinforcement is defined as when an event, occurs due to a particular behavior, increases
the strength and the frequency of the behavior. In other words, it has a positive effect on behavior.
2. Negative –
Negative Reinforcement is defined as strengthening of behavior because a negative condition
is stopped or avoided.

Various Practical applications of Reinforcement Learning –

 RL can be used in robotics for industrial automation.

 RL can be used in machine learning and data processing
 RL can be used to create training systems that provide custom instruction and materials according
to the requirement of students.

Faculty Name : Mrs Swapna Subject Name :ML

 1. Robotics:
a. RL is used in Robot navigation, Robo-soccer, walking, juggling, etc.
Control:
. RL can be used for adaptive control such as Factory processes, admission control in
telecommunication, and Helicopter pilot is an example of reinforcement learning.
Game Playing:
. RL can be used in Game playing such as tic-tac-toe, chess, etc.
Chemistry:
. RL can be used for optimizing the chemical reactions.
Business:
Faculty Name : Mrs Swapna Subject Name :ML
 . RL is now used for business strategy planning.
Manufacturing:
. In various automobile manufacturing companies, the robots use deep reinforcement learning to pick
goods and put them in some containers.
Finance Sector:
. The RL is currently used in the finance sector for evaluating trading strategies.

Terms used in Reinforcement Learning

o Agent(): An entity that can perceive/explore the environment and act upon it.
o Environment(): A situation in which an agent is present or surrounded by. In RL, we assume
the stochastic environment, which means it is random in nature.
o Action(): Actions are the moves taken by an agent within the environment.
o State(): State is a situation returned by the environment after each action taken by the agent.
o Reward(): A feedback returned to the agent from the environment to evaluate the action of the
agent.
o Policy(): Policy is a strategy applied by the agent for the next action based on the current state.
o Value(): It is expected long-term retuned with the discount factor and opposite to the short-term
reward.
o Q-value(): It is mostly similar to the value, but it takes one additional parameter as a current
action (a).

From the above discussion, we can say that Reinforcement Learning is one of the most interesting and
useful parts of Machine learning. In RL, the agent explores the environment by exploring it without any
human intervention. It is the main learning algorithm that is used in Artificial Intelligence. But there are
some cases where it should not be used, such as if you have enough data to solve the problem, then other
ML algorithms can be used more efficiently. The main issue with the RL algorithm is that some of the
parameters may affect the speed of the learning, such as delayed feedback.

Q-learning OR Quality learning:

is a model-free reinforcement learning algorithm to learn the value of an action in a particular state.
... "Q" refers to the function that the algorithm computes – the expected rewards for an action taken in
a given state

Faculty Name : Mrs Swapna Subject Name :ML

What is q-learning?

Q-learning is an off policy reinforcement learning algorithm that seeks to find the best action to take
given the current state. It’s considered off-policy because the q-learning function learns from actions
that are outside the current policy, like taking random actions, and therefore a policy isn’t needed.
More specifically, q-learning seeks to learn a policy that maximizes the total reward.

What’s ‘Q’?

The ‘q’ in q-learning stands for quality. Quality in this case represents how useful a given action is in
gaining some future reward.

Create a q-table

When q-learning is performed we create what’s called a q-table or matrix that follows the shape
of [state, action] and we initialize our values to zero. We then update and store our q-values after
an episode.

Q-learning and making updates

The next step is simply for the agent to interact with the environment and make updates to the state
action pairs in our q-table Q[state, action].

Taking Action: Explore or Exploit

An agent interacts with the environment in 1 of 2 ways. The first is to use the q-table as a reference
and view all possible actions for a given state. The agent then selects the action based on the max
value of those actions. This is known as exploiting since we use the information we have available
to us to make a decision.

The second way to take action is to act randomly. This is called exploring. Instead of selecting
actions based on the max future reward we select an action at random. Acting randomly is
important because it allows the agent to explore and discover new states that otherwise may not be
selected during the exploitation process.

Faculty Name : Mrs Swapna Subject Name :ML

Q-Values or Action-Values: Q-values are defined for states and actions. Q(S,A) is an estimation of
how good is it to take the action A at the state S. This estimation of Q(S,A) will be iteratively computed
using the TD- Update rule
Rewards and Episodes: An agent over the course of its lifetime starts from a start state, makes a
number of transitions from its current state to a next state based on its choice of action and also the
environment the agent is interacting in. At every step of transition, the agent from a state takes an action,
observes a reward from the environment, and then transits to another state. If at any point of time the
agent ends up in one of the terminating states that means there are no further transition possible. This is
said to be the completion of an episode.

Temporal Difference or TD-Update:

The Temporal Difference or TD-Update rule can be represented as follows

This update rule to estimate the value of Q is applied at every time step of the agents interaction with the
environment. The terms used are explained below. :

S: Current State of the agent.

A: Current Action Picked according to some policy.

S‟ : Next State where the agent ends up.

A‟ : Next best action to be picked using current Q-value estimation, i.e. pick the action with the
maximum Q-value in the next state.
R : Current Reward observed from the environment in Response of current action.
GAMMA=(>0 and <=1) : Discounting Factor for Future Rewards. Future rewards are less valuable than
current rewards so they must be discounted. Since Q-value is an estimation of expected rewards from a
state, discounting rule applies here as well.
ALPHA : Step length taken to update the estimation of Q(S, A).

Faculty Name : Mrs Swapna Subject Name :ML

This is one of the form of Learning method in which the agent moves to one state and other state …and
predicts the reward expected

It is a supervised learning process in which the training signal for a prediction is a future
prediction. TD algorithms are often used in reinforcement learning to predict a measure of the total
amount of reward expected over the future.

TD Learning focuses on predicting a variable's future value in a sequence of states. Temporal difference
learning was a major breakthrough in solving the problem of reward prediction. You could say that iIt
employs a mathematical trick that allows it to replace complicated reasoning with a simple learning
procedure that can be used to generate the very same results.
The trick is that rather than attempting to calculate the total future reward, temporal difference learning
just attempts to predict the combination of immediate reward and its own reward prediction at the next
moment in time. Now when the next moment comes and brings fresh information with it, the new
prediction is compared with the expected prediction. If these two predictions are different from each
other, the TD algorithm will calculate how different the predictions are from each other and make use of
this temporal difference to adjust the old prediction toward the new prediction.

Faculty Name : Mrs Swapna Subject Name :ML

The temporal difference algorithm always aims to bring the expected prediction and the new prediction
together, thus matching expectations with reality and gradually increasing the accuracy of the entire
chain of prediction.
Temporal Difference Learning aims to predict a combination of the immediate reward and its own
reward prediction at the next moment in time.
In TD Learning, the training signal for a prediction is a future prediction. temporal difference methods
tend to adjust predictions to match later, more accurate, predictions for the future, much before the final
outcome is clear and know. This is essentially a type of bootstrapping.
For a action to take given a particular state. The distribution π(a∣s) is used for a stochastic policy and
a mapping function π:S→A is used for a deterministic policy(fixed policy or rule ) and non
deterministic (customized policy with time bound) where S is the set of possible states and A is the set
of possible actions.

Dynamic Programming

Dynamic programming algorithms solve a category of problems called planning problems. Herein given
the complete model and specifications of the environment , we can successfully find an optimal policy
for the agent to follow. It contains two main steps:

1. Break the problem into subproblems and solve it

2. Solutions to subproblems are cached or stored for reuse to find overall optimal
solution to the problem at hand

• Dynamic programming is a method for solving complex problems by breaking them down into
sub-problems. The solutions to the sub-problems are combined to solve overall problem.

MDP is used in Dynamic programming to solve complex task

Faculty Name : Mrs Swapna Subject Name :ML
Faculty Name : Mrs Swapna Subject Name :ML
Faculty Name : Mrs Swapna Subject Name :ML
 UNIT 5

Analytical Learning-1

Inductive learning methods such as neural network and decision tree learning require a certain number of training
examples to achieve a given level of accuracy.

Analytical learning uses prior knowledge and deductive reasoning to augment the information provided by the
training examples. This chapter considers an analytical learning method called explanation-based learning (EBL)

In explanation-based learning, prior knowledge is used to analyze, or explain, how each observed training example
satisfies the target concept.

This explanation is then used to distinguish the relevant features of the training example from the irrelevant features
Explanation-based learning has been successfully applied to learning search control rules for a variety of planning
and scheduling tasks.

Inductive Learning methods: that is, methods that generalize from observed training examples by identifying
features that empirically distinguish positive from negative training examples.

Decision tree learning, neural network learning, inductive logic programming, and genetic algorithms are all
examples of inductive methods that operate in this fashion. The key practical limit on these inductive learners is that
they perform poorly when insufficient data is available.

Explanation-based learning is one such approach. It uses prior knowledge to analyze, or explain, each training
example in order to infer which example features are relevant to the target function and which are irrelevant. These
explanations enable it to generalize more accurately than inductive systems that rely on the data.

Explanation based learning uses prior knowledge to reduce the complexity of the hypothesis space to be searched,
thereby reducing sample complexity and improving generalization accuracy of the learner it is supported with
evidence called domain theory(it is refers to the evidence that supports the prior data ) .

EXAMPLE OF EXPLANATION: Chess Game

Faculty Name : Mrs Swapna Subject Name :ML

 Inductive and Analytical Learning Problems

Faculty Name : Mrs Swapna Subject Name :ML

 Let us introduce in detail a second example of an analytical learning problem--one that we will use for illustration
throughout this chapter. Consider an instance space X in which each instance is a pair of physical objects. Each of
the two physical objects in the instance is described by the predicates Color, Volume, Owner, Material, Type, and
Density, and the relationship between the two objects is described by the predicate On.

Given this instance space, the task is to learn the target concept "pairs of physical objects, such that one can be
stacked safely on the other," denoted by the predicate SafeToStack(x,y). Learning this target concept might be
useful, for example, to a robot system that has the task of storing various physical objects within a limited
workspace. The full definition of this analytical learning task is given in

Table 11.1.

Faculty Name : Mrs Swapna Subject Name :ML

 Horn clauses express a subset of statements of first-order logic.
Horn clause supports atleast one positive and negative literal representation
A Horn Clause is a clause with at most one positive literal, it is thus either: A single positive literal, which is
regarded as a fact, One or more negative literals, with no positive literal.

Faculty Name : Mrs Swapna Subject Name :ML

 ANALYZE THE EXPLANATION

Faculty Name : Mrs Swapna Subject Name :ML

 Learning With Perfect Domain Theories: PROLOG-EBG-PROGRAM LOGIC

This section presents an algorithm called PROLOG-EBG (Kedar-Cabelli and McCarty 1987) that is representative
of several explanation-based learning algorithms.

PROLOG-EBG is a sequential covering algorithm. Prolog stands for programming in logic. it is a logic
programming language for artificial intelligence. An artificial intelligence developed in Prolog will examine the
link between a fact, a true statement, and a rule, a conditional statement, in order to come up with a question, or end
objective.

PROLOG-EBG is guaranteed to output a hypothesis (set of rules) that is itself correct and that covers the observed
positive training examples. For any set of training examples, the hypothesis output by PROLOG-EBG constitutes a
set of logically sufficient conditions for the target concept, according to the domain theory.

A domain theory is said to be correct if each of its assertions is a truthful statement about the world.

A domain theory is said to be complete with respect to a given target concept and instance space, if the domain
theory covers every positive example in the instance space.

Faculty Name : Mrs Swapna Subject Name :ML

 PROLOGEBG computes the most general rule that can be justified by the explanation, by computing the weakest
preimage of the explanation

PROLOG-EBG computes the weakest preimage of the target concept with respect to the explanation, using a
general procedure called regression (Waldinger 1977). The regression procedure operates on a domain theory
represented by an arbitrary set of Horn clauses.

It works iteratively backward through the explanation, first computing the weakest preimage of the target concept
with respect to the final proof step in the explanation, then computing the weakest preimage of the resulting
expressions with respect to the preceding step, and so on.

The procedure terminates when it has iterated over all steps :in the explanation, yielding the weakest precondition
of the target concept with respect to the literals at the leaf nodes of the explanation

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
 In prolog, We declare some facts. These facts constitute the Knowledge Base of the system. We can query against
the Knowledge Base. We get output as affirmative if our query is already in the knowledge Base or it is implied
by Knowledge Base, otherwise we get output as negative. So, Knowledge Base can be considered similar to
database, against which we can query. Prolog facts are expressed in definite pattern. Facts contain entities and
their relation. Entities are written within the parenthesis separated by comma (, ). Their relation is expressed at
the start and outside the parenthesis. Every fact/rule ends with a dot (.). So, a typical prolog fact goes as follows :
Format : relation(entity1, entity2, ....k'th entity).

Example :
friends(raju, mahesh).
singer(sonu).
odd_number(5).

Explanation :
These facts can be interpreted as :

Faculty Name : Mrs Swapna Subject Name :ML

raju and mahesh are friends.
sonu is a singer.
5 is an odd number.

Key Features of Prolog used in AI:

1. Unification : The basic idea is, can the given terms be made to represent the same structure.
2. Backtracking : When a task fails, prolog traces backwards and tries to satisfy previous task.
3. Recursion : Recursion is the basis for any search in program.

Running queries :

A typical prolog query can be asked as :

Query 1 : ?- singer(sonu).
Output : Yes.

Explanation : As our knowledge base contains

the above fact, so output was 'Yes', otherwise
it would have been 'No'.

Query 2 : ?- odd_number(7).
Output : No.

Explanation : As our knowledge base does not

contain the above fact, so output was 'No'.

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
 Thus, in its pure form EBL involves reformulating the domain theory to produce general rules that classify
examples in a single inference step.

This kind of knowledge reformulation is sometimes referred to as knowledge compilation, indicating that the
transformation is an efficiency improving one

EXPLANATION-BASED LEARNING OF SEARCH CONTROL KNOWLEDGE

PRODIGY and SOAR demonstrate that explanation-based learning methods can be successfully applied to acquire
search control knowledge in a variety of problem domains.

Prodigy means-ends planning strategy.

The PRODIGY architecture was initially conceived by Jaime Carbonell and Steven Minton, as an Artificial
Intelligence (AI) system to test and develop ideas on the role of machine learning in planning and problem solving.

In general, learning in problem solving seemed meaningless without measurable performance improvements.

Thus, PRODIGY was created to be a testbed for the systematic investigation of the loop between learning and
performance in planning systems.

As a result, PRODIGY consists of a core general- purpose planner and several learning modules that refine both the
planning domain knowledge and the control knowledge to guide the search process effectively

A planning problem is defined by

(1) a set of available objects of each type,

(2) an initial state , and

Faculty Name : Mrs Swapna Subject Name :ML

 (3) a goal statement .

The Algorithm
Table 4 shows the basic procedure to learn quality-enhancing control knowledge, in the case that a
human expert provides a better plan. Steps 2, 3 and 4 correspond to the interactive plan checking
module, that asks the expert for a better solution and checks for its correctness. Step 6 constructs a
problem solving trace from the expert solution and obtains decision points where control knowledge is
needed, which in turn become learning opportunities. Step 8 corresponds to the actual learning phase. It
compares the plan trees obtained from the problem solving traces in Step 7, explains why one solution
was better than the other, and builds new control knowledge. These steps are described now in detail.

Faculty Name : Mrs Swapna Subject Name :ML

 1. Run PRODIGY with the current set of control rules and obtain a solution
.
2. Show to the expert.
Expert provides new solution possibly using as a guide.

3. Test . If it solves the problem, continue. Else go back to step 2.

4. Apply the plan quality evaluation function to .
If it is better than , continue. Else go back to step 2.

5. Compute the partial order for identifying the goal dependencies between
plan steps.
6. Construct a problem solving trace corresponding to a solution that satisfies
.
This determines the set of decision points in the problem solving trace where
control knowledge is missing.

7. Build the plan trees and , corresponding respectively to the search trees for and
.
8. Compare and explaining why is better than , and build control rules.

Table 4: Top level procedure to learn quality-enhancing control knowledge.

A second example of a general problem-solving architecture that incorporates a form of

explanation-based learning is the SOAR system (Laird et al. 1986; Newel1 1990). SOAR
supports a broad variety of problem-solving strategies that subsumes PRODIGY'S means-
ends planning strategy.
Like PRODIGY, however, SOAR learns by explaining situations in which its current
search strategy leads to inefficiencies.
SOAR has been applied in a great number of problem domains and has also been proposed
as a psychologically plausible model of human learning processes (Newel1 1990).

SOAR (security orchestration, automation and response) is a stack of compatible software

programs that enables an organization to collect data about security threats and respond

Faculty Name : Mrs Swapna Subject Name :ML

to security events without human assistance. The goal of using a SOAR platform is to improve
the efficiency of physical and digital security operations.

What is SOAR?

SOAR platforms have three main components: security orchestration, security automation and security response.

Faculty Name : Mrs Swapna Subject Name :ML

 Benefits of SOAR
SOAR platforms offer many benefits for enterprise security operations (SecOps) teams,
including the following:

 Faster incident detection and reaction times. The volume and velocity of security threats
and events are constantly increasing. SOAR's improved data context, combined with
automation, can bring lower mean time to detect (MTTD) and mean time to respond (MTTR).
By detecting and responding to threats more quickly, their impact can be lessened.

 Better threat context. By integrating more data from a wider array of tools and systems,
SOAR platforms can offer more context, better analysis and up-to-date threat information.

 Simplified management. SOAR platforms consolidate various security systems' dashboards

into a single interface. This helps SecOps and other teams by centralizing information and data
handling, simplifying management and saving time.

 Scalability. Scaling time-consuming manual processes can be a drain on employees and even
impossible to keep up with as security event volume grows. SOAR's orchestration, automation
and workflows can meet scalability demands more easily.

 Boosting analysts' productivity. Automating lower-level threats augments SecOps and

security operations center (SOC) teams' responsibilities, enabling them to prioritize tasks more
effectively and respond to threats that require human intervention more quickly.

Faculty Name : Mrs Swapna Subject Name :ML

 Analytical Learning-2-

Faculty Name : Mrs Swapna Subject Name :ML

Motivation for combining Inductive and Analytical approaches which address following

specific properties

1. Given no domain theory, it should learn at least as effectively as purely inductive methods.
2. Given a perfect domain theory, it should learn at least as effectively as purely analytical methods.
3. Given an imperfect domain theory and imperfect training data, it should combine the two to out
perform either purely inductive or purely analytical methods.
4. It should accommodate an unknown level of error in the training data.
Faculty Name : Mrs Swapna Subject Name :ML
 5. It should accommodate an unknown level of error in the domain theory
INDUCTIVE-ANALYTICAL APPROACHES TO LEARNING

The Learning Problem specified as

To summarize, the learning problem considered in this chapter is

Given:

A set of training examples D, possibly containing errors

A domain theory B, possibly containing errors

A space of candidate hypotheses H

Determine:

A hypothesis that best fits the training examples and domain theory

To address this learning problem we develop a hypothesis space search combining both
inductive and analytical approaches

we explore three different methods for using prior knowledge to alter the search performed by purely
inductive methods.

1. USE PRIOR KNOWLEDGE TO DERIVE AN INITIAL HYPOTHESIS FROM WHICH TO BEGIN THE SEARCH

2. USE PRIOR KNOWLEDGE TO ALTER THE OBJECTIVE OF THE HYPOTHESIS SPACE SEARCH.

3. USING PRIOR KNOWLEDGE TO AUGMENT SEARCH STEPS

1. Use prior knowledge to derive an initial hypothesis from which to begin the search

In this approach the domain theory B is used to construct an initial hypothesis ho that is consistent with
B. A standard inductive method is then applied, starting with the initial hypothesis ho.

This approach is used by the KBANN (Knowledge-Based Artificial Neural Network) algorithm to learn
artificial neural networks.

In KBANN an initial network is first constructed so that for every possible instance, the classification
assigned by the network is identical to that assigned by the domain theory.

Faculty Name : Mrs Swapna Subject Name :ML

The BACKPROPAGATION algorithm is then employed to adjust the weights of this initial network as
needed to fit the training examples. It is easy to see the motivation for this technique: if the domain
theory is correct, the initial hypothesis will correctly classify all the training examples and there will be
no need to revise it.

However, if the initial hypothesis is found to imperfectly classify the training examples, then it will be
refined inductively to improve its fit to the training examples.

The KBANN Algorithm

The KBANN algorithm exemplifies the initialize-the-hypothesis approach to using domain theories.

The input and output of KBANN are as follows:

Faculty Name : Mrs Swapna Subject Name :ML

An Illustrative Example

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
Remarks To summarize, KBANN analytically creates a network equivalent to the given domain theory,
then inductively refines this initial hypothesis to better fit the training data.

In doing so, it modifies the network weights using backpropgation as needed to overcome
inconsistencies between the domain theory and observed data.

Limitations of KBANN include the fact that it can accommodate only propositional domain theories

It is also possible for KBANN to be misled when given highly inaccurate domain theories, so that its
generalization accuracy can deteriorate below the level of BACKPROPAGATION

2.USING PRIOR KNOWLEDGE TO ALTER THE SEARCH OBJECTIVE

Faculty Name : Mrs Swapna Subject Name :ML

In this approach, the goal criterion G is modified to require that the output hypothesis fits the domain
theory as well as the training examples.

For example, the EBNN(EXPLANATION BASED NEURAL NETWORK) system described below learns
neural networks in this way. Whereas inductive learning of neural networks performs gradient descent
search to minimize the squared error of the network over the training data, EBNN performs gradient
descent to optimize a different criterion. This modified criterion includes an additional term that
measures the error of the learned network relative to the domain theory.

The TANGENTPROP Algorithm TANGENTPROP (Simard et al. 1992) accommodates domain knowledge
expressed as derivatives of the target function with respect to transformations of its inputs. Consider a
learning task involving an instance space X and target function f.

The TANGENTPROP algorithm assumes various training derivatives of the target function are also
provided. For example, if each instance xi is described by a single real value, then each training example
may be of the form (xi, f (xi), q lx, ). Here lx, denotes the derivative of the target function f with respect
to x, evaluated at the point x = xi.

To develop an intuition for the benefits of providing training derivatives as well as training values during
learning, consider the simple learning task depicted in Figure

The task is to learn the target function f shown in the leftmost plot of the figure, based on the three
training examples shown: (xl, f (xl)), (x2, f (x2)), and (xg, f (xg)).

Given these three training examples, the BACKPROPAGATION algorithm can be expected to hypothesize
a smooth function, such as the function g depicted in the middle plot of the figure. The rightmost plot
shows the effect of

providing training derivatives, or slopes, as additional information for each training example (e.g., (XI, f
(XI), I,, )). By fitting both the training values f (xi) and these training derivatives PI,, the learner has a
better chance to correctly generalize from the sparse training data.

To summarize, the impact of including the training derivatives is to override the usual syntactic inductive
bias of BACKPROPAGATION that favors a smooth interpolation between points, replacing it by explicit
input information about required derivatives. The resulting hypothesis h shown in the rightmost plot of
the figure provides a much more accurate estimate of the true target function f.

Faculty Name : Mrs Swapna Subject Name :ML

Each transformation must be of the form sj(a, x) where aj is a continuous parameter, where sj is
differentiable, and where sj(O, x) = x (e.g., for rotation of zero degrees the transformation is the identity
function). For each such transformation, sj(a, x),

In the Figure one f(X) are the hypothesis and x1 , x2 ,x3 are the instances and these instances fit to
proper hypothesis shown in first figure and in second fig we can see the instances classified and
machine learns to fit to proper hypothesis by doing necessary modification by using

TANGEPROP considers the squared error between the specified training derivative and the actual
derivative of the learned neural network. The modified error function is

where p is a constant provided by the user to determine the relative importance of fitting training values
versus fitting training derivatives.

Notice the first term in this definition of E is the original squared error of the network versus training
values, and the second term is the squared error in the network versus training derivatives.

In the third figure we can see the instances are classified properly and maintaining accuracy.

An Illustrative Example

Faculty Name : Mrs Swapna Subject Name :ML

Remarks To summarize, TANGENTPROP uses prior knowledge in the form of desired derivatives of the
target function with respect to transformations of its inputs.

It combines this prior knowledge with observed training data, by minimizing an objective function that
measures both the network's error with respect to the training example values (fitting the data) and its
error with respect to the desired derivatives (fitting the prior knowledge).

Faculty Name : Mrs Swapna Subject Name :ML

It is interesting to compare the search through hypothesis space (weight space) performed by
TANGENTPROP, KBANN, and BACKPROPAGATION.

TANGENTPROP incorporates prior knowledge to influence the hypothesis search by altering the
objective function to be minimized by gradient descent

TANGENTPROP objective will be a subset of those satisfying the weaker BACKPROPAGATION objective.
The difference between these two sets of final hypotheses is the set of incorrect hypotheses that will be
considered by BACKPROPAGATION, but ruled out by TANGENTPROP due to its prior knowledge.

The EBNN Algorithm

The EBNN (Explanation-Based Neural Network learning) algorithm (Mitchell and Thrun 1993a; Thrun
1996) builds on the TANGENTPROP algorithm in two significant ways.

First, instead of relying on the user to provide training derivatives, EBNN computes training derivatives
itself for each observed training example. These training derivatives are calculated by explaining each
training example in terms of a given domain theory, then extracting training derivatives from this
explanation.

Faculty Name : Mrs Swapna Subject Name :ML

Second, EBNN addresses the issue of how to weight the relative importance of the inductive and
analytical components of learning .

The inputs to EBNN include (1) a set of training examples of the form (xi, f (xi)) with no training
derivatives provided, and (2) a domain theory analogous to that used in explanation-based learning and
in KBANN, but represented by a set of previously trained neural networks The output of EBNN is a new
neural network that approximates the target function f. This learned network is trained to fit both the
training examples (xi, f (xi)) and training derivatives of f extracted from the domain theory. Fitting the
training examples (xi, f (xi)) constitutes the inductive component of learning, whereas fitting the training
derivatives extracted from the domain theory provides the analytical component.

Faculty Name : Mrs Swapna Subject Name :ML

Remarks To summarize, the EBNN algorithm uses a domain theory expressed as a set of previously
learned neural networks, together with a set of training examples, to train its output hypothesis (the
target network).

For each training example EBNN uses its domain theory to explain the example, then extracts training
derivatives from this explanation.

For each attribute of the instance, a training derivative is computed that describes how the target
function value is influenced by a small change to this attribute value, according to the domain theory.

USING PRIOR KNOWLEDGE TO AUGMENT SEARCH STEPS

The two previous sections examined two different roles for prior knowledge in learning: initializing the
learner's hypothesis and altering the objective function that guides search through the hypothesis
space.

In this section we consider a third way of using prior knowledge to alter the hypothesis space search:
using it to alter the set of operators that define legal steps in the search through the hypothesis space.
This approach is followed by systems such as FOCL

The First Order Combined Learner (FOCL) Algorithm is an extension of the purely inductive, FOIL
Algorithm. It uses domain theory to further improve the search for the best-rule and greatly improves
accuracy.
First Order Inductive Learner (FOIL)
In machine learning, (FOIL) is a rule-based learning algorithm. It is a natural extension of SEQUENTIAL-
COVERING and LEARN-ONE-RULE algorithms

FOCL also tends to perform an iterative process of learning a set of best-rules to cover the
training examples and then remove all the training examples covered by that best rule. (using a
sequential covering algorithm)
However, what makes the FOCL algorithm more powerful is the approach that it adapts while
searching for that best-rule.

Faculty Name : Mrs Swapna Subject Name :ML

Faculty Name : Mrs Swapna Subject Name :ML
At each point in it moves from general-to-specific search, FOCL expands its current hypothesis h using
the following two operators

1. For each operational literal that is not part of h, create a specialization of h by adding this single literal
to the precondition s. This is also the method used by FOIL to generate candidate successors. The solid
arrows in Figure 12.8 denote this type of specialization

2. Create an operational, logically sufficient condition for the target concept according to the domain
theory. Add this set of literals to the current preconditions of h.

Finally, prune the preconditions of h by removing any literals that are unnecessary according to the
training data.

Faculty Name : Mrs Swapna Subject Name :ML