ML Question Bank
ML Question Bank
ML Question Bank
(Autonomous)
Dundigal, Hyderabad - 500 043
COURSE OBJECTIVES:
The students will try to learn:
COURSE OUTCOMES:
After successful completion of the course, students should be able to:
CO 1 Recognize the characteristics of machine learning that make it Remember
useful to real-world problem.
CO 2 Discuss the steps involved in designing a learning system by Understand
consideringwell posed learning problems.
CO 3 Analyze the underlying mathematical relationships within and Analyze
across machine learning algorithms.
CO 4 Relate he hypothesis space searchfor an application using decision Remember
treelearning.
CO 5 Explain models for reasoning with uncertainty as well as the use Understand
of unreliable information.
CO 6 Identify appropriate learning functions as activation function for Understand
neural network design.
CO 7 Make use of optimizing function for reducing the error in Apply
prediction model.
CO 8 Develop artificial neural networkusing back propagation algorithm Apply
for different applications.
CO 9 Demonstrate Naı̈ve Bayes algorithm based on Bayes theorem for Apply
classification problem.
CO 10 Make Use of k-nearest neighbor algorithm for solving both Apply
classification and regression problems.
CO 11 Discuss he reinforcement learning by observing the current Understand
environmentstate.
CO 12 Identify appropriate machine learning techniques and computing Understand
environment suitable for the applications.
QUESTION BANK:
MODULE I
TYPES OF MACHINE LEARNING
PART-A PROBLEM SOLVING AND CRITICAL THINKING QUESTIONS
Q.No QUESTION Taxonomy How does this subsume CO’s
the level
1 Mention three computer Understand The learner to recall the CO1
applications for which concept of machine learning
machine learning approaches and also able to
approaches seem identify the applications
appropriate and three for which are in need of
which they seem machine learning
inappropriate. approaches.
2 Let consider designing a Apply The learner to recall the CO2
program to learn play checkers learning program
checkers, with the goal of and mention the
entering it in the world appropriate target function
checkers tournament. and its representation.
Propose a target function to
be learned and a target
representation.
Page 2
3 For the given dataset, derive Apply The learner to recall the CO4
the hypothesis for the Enjoy Find-S algorithm and
Sport task by using Find-S should appropriately derive
algorithm. hypothesis for a problem
domain.
6 For the given dataset, derive Analyze The learner to recall the CO4
the set of hypothesis for the consistent hypothesis and
Enjoy Sport task by using candidate elimination
find S algoritham. algorithm concepts and try
to find out whether the
obtained hypothesis is
consistent or not.
7 For the given dataset, derive Analyze The learner to recall the CO4
the set of hypothesis for the consistent hypothesis and
enjoy sport task by using candidate elimination
list then elimination algorithm concepts and try
algoritham. to find out whether the
obtained hypothesis is
consistent or not.
Page 3
9 Explain more general than Understand
The learner to recall the CO3
or equal to relation in terms concept of ordering of
of sets of instances that hypothesis in order to
satisfy any two hypothesis. mention relationship
between hypotheses.
10 Explain the biased Understand The learner to recall the CO4
hypothesis space in bias concept and how it is
candidate elimination included in obtaining
algorithm version space for a problem
domain.
PART B-LONG ANSWER QUESTIONS
1 Describe the Enjoy Sport Understand The learner to mention CO4
concept learning task and concept learning and with
hypothesis space of Enjoy the help of the example of
sport. Enjoy Sport and also the
hypothesis space concerned
with the task.
2 Explain Find-S algorithm Understand The learner to explain CO12
with respect to Enjoy Sport Find-S algorithm and with
concept learning task. the help of the example of
Enjoy Sport task.
3 Explain in detail the issues Understand The learner to recall the CO5
in Find-S algorithm. Find-S algorithm and
should mention all issues in
that.
4 Describe Understand The learner to mention CO5
List-Then-Eliminate List-Then-Eliminate
algorithm and its drawback. algorithm and should
mention all drawbacks in
that.
5 Explain version space and Understand The learner to describe CO5
also define general and version space and how
specific boundary with an general and specific
example. boundaries are related to it.
6 Distinguish data mining and Understand The learner should know CO1
machine learning. the limitations of data
mining to differentiate with
machine learning.
7 Compare artificial Analyze The learner to mention the CO1
intelligence, machine purpose of these three
learning and deep learning disciplines and the way they
with their relationship. are inter-related.
Page 4
8 Explain different algorithm Understand The learner to recall the CO1
approaches in machine various ways the machine
learning learning techniques are
present and explain the
ways they are implemented.
9 Explain inductive bias in Understand The learner to recall the CO5
candidate elimination concept of inductive bias
algorithm. and candidate elimination
algorithm then should
describe how inductive bias
is present in this algorithm.
10 Explain the limitations of Understand The learner to mention CO1
version space learned by version space in candidate
candidate elimination elimination algorithm and
algorithm. the drawbacks present in its
hypothesis.
11 Describe the final version Understand The learner to derive CO12
space of Enjoy Sport task version space for enjoy sport
obtained from candidate task using candidate
elimination algorithm with elimination algorithm.
a neat diagram.
12 Explain in detail issues in Understand The learner to recall CO1
machine learning. machine learning and the
issues in it related to
training experience and
generalization of target
function.
13 State the choices in learning Remember —- CO2
the checkers learning
program with the help of a
diagram.
14 Explain the steps in Understand The learner to explain the CO2
designing a learning system. five important steps that
are to be followed when
developing a machine
learning model.
15 State the representation of Understand The learner should have an CO2
target function of the idea of target function in
checkers learning program over to mathematically
and explain each term in it. represent it for the checkers
learning problem.
Page 5
16 Explain LMS weight update Understand The learner should possess CO3
rule in adjusting weights. the knowledge about
function approximation, in
this process the way in
which weights are adjusted.
17 Describe the final design of Understand The learner should know CO2
the checkers learning system the Software development
with a proper diagram. process of machine learning
algorithms.
18 Compare Find-S and Analyze The learner to recall both CO5
Candidate Elimination these algorithms and their
Algorithms. applicability and drawbacks.
19 Explain the steps in Understand The learner to recall the CO12
Candidate Elimination steps in candidate
Algorithm. elimination algorithm in
order to properly obtain the
version space.
20 State the rule for estimating Apply The learner to recall the CO2
training values for checkers function approximation step
learning program during in designing a learning
function approximation. problem and mention the
rule of function
approximation for a
checkers problem.
PART C-SHORT ANSWER QUESTIONS
1 Explain General-to-Specific Understand The learner first should CO4
Ordering of Hypotheses have an idea of hypothesis
with an example. and then decides the
ordering of hypothesis.
2 Define hypothesis in Remember —- CO4
machine learning.
3 State the general notation Remember —- CO4
of hypothesis in machine
learning with positive
example and negative
example.
4 Explain the use of Find-S Understand The learner should know CO2
algorithm and which find s algorithm, then its
hypothesis is considered in use and the hypothesis
this algorithm? considered are explained.
5 Define inductive learning in Remember —- CO1
machine learning.
6 Define machine learning Remember —- CO1
according to Tom Mitchell.
Page 6
7 Explain task, experience Understand The learner should know CO2
and performance in checkers how to well pose a machine
learning problem. learning problem to explain
task, experience and
performance.
8 Mention any four domains Remember —- CO1
where machine learning
applications are useful.
9 Define a well posed learning Remember —- CO2
problem.
10 Explain the importance of Understand The learner should CO1
Machine Learning. understand the difficulties
in traditional techniques in
order to find the importance
of machine learning in the
problem domain.
11 List a fewpopular examples Remember —- CO1
of machine learning.
12 Draw the structure of Remember —- CO2
learning problem.
13 Distinguish traditional and Understand The learner to define CO1
machine learning program. traditional and machine
learning program in terms
of input, processing and
output.
14 Compare predictive and Analysis The learner to define
descriptive learning task. predictive and descriptive
tasks and how they are
varied.
15 State positive and negative Remember —- CO1
example in training data
set?
16 List the steps in designing a Remember —- CO2
learning system.
17 Statethe perspective of Remember —- CO1
machine learning.
18 Define version space in Remember —- CO1
machine learning.
19 Explain when you can say Understand The learner to define how CO4
that the hypothesis is hypothesis is obtained and
consistent? when it becomes consistent.
Page 7
20 Describe general boundary Understand
The learner to define CO4
and specific boundary with general and specific
respect to hypothesis space hypothesis in the predefined
and training data. hypothesis space.
MODULE II
DECISION TREE LEARNING
PART-A PROBLEM SOLVING AND CRITICAL THINKING QUESTIONS
1 For example, suppose S is a Apply The learner to recall the CO4
collection of information gain measure.
training-example days
described by attributes
including Wind, which can
have the values Weak or
Strong. As before, assume S
is a collection containing 14
examples, [9+, 5-]. Of these
14 examples, suppose 6 of
the positive and 2 of the
negative examples have
Wind = Weak, andthe
remainder have Wind =
Strong. Calculate
information gain due to
sorting the original 14
Examples by the attribute
Wind.
2 Calculate the entropy of Apply The learner to recall the CO4
each attribute for the given entropy measure in order to
dataset. obtain best hypothesis.
Page 8
3 Find the attribute that is Apply The learner to recall the CO4
best suited for the root information gain measure in
node for the given dataset. order to obtain best
hypothesis.
4 Find the attribute that is Apply The learner to recall the CO4
best suited for the root information gain measure in
node for the given dataset, order to obtain best
outlook of third attribute hypothesis.
value .
6 What do you infer from the Analyze The learner to relate the CO5
below diagram of a training concepts of overfitting and
data set. underfitting in terms of
training data.
7 Explain how high bias and Understand The learner to figure out the CO5
low variance contribute to relation between bias and
underfitting. variance.
Page 9
8 Explain how low bias and Understand The learner to figure out the CO5
high variance contribute to relation between bias and
overfitting. variance.
9 Derive with a class dataset Apply The learner to recall the CO4
with marks, overall process of decision tree
participation and discipline construction.
attributes a best hypothesis
using decision tree learning.
10 Construct the decision tree Apply The learner to recall the CO4
for the given data set by process of decision tree
considering information gain construction.
at each level.
Page 10
6 Explain hypothesis space Understand The learner to explain how CO4
search in decision tree the best hypothesis is
learning. inferred from a hypothesis
space.
7 List the steps in ID3 Remember —- CO4
algorithm with proper
explanation.
8 Mention the appropriate Remember —- CO5
problems for decision tree
learning.
9 Demonstrate decision tree Remember —- CO12
representation with the help
of a diagram.
10 State the terminating Remember —- CO4
conditions of decision tree
learning algorithm.
11 Which type of inference is Analyze The learner to recall CO4
incorporated in decision tree difference inferences and to
learning model? Justify mention the proper
your answer. inference that suits in
decision tree learning.
12 List the different issues in Remember —- CO5
decision tree learning
technique?
13 Explain incorporating Understand The learner to address the CO5
continuous valued attributes issues related to decision
in Decision tree learning tree learning.
with an example.
14 Describe the need to Understand The learner to identify the CO5
convert the decision tree to need to change decision
rules before pruning? trees to rules in terms of
processing.
15 Explain with an example Understand The learner to relate CO5
rule antecedent and rule precondition and post
consequent in rule post condition in context of rule
pruning. post pruning.
16 Explain different formulae Understand The learner to recall all the CO4
for attribute selection attribute selection measures
measure instead of for handling issue related to
information gain attribute cost.
for handling attributes with
differing costs in decision
trees.
Page 11
17 Describe the method of Understand The learner to handle the CO5
handling missing attribute issues in decision tree
value in decision tree learning approach.
learning using probability.
18 Explain alternative Understand The learner to recall all the CO4
measures for selecting attribute selection measures
attributes other than for handling issues.
information gain.
19 State noise in the training Remember —- CO5
data. Mention the issues
related to it.
20 a) Demonstrate the Analyze The learner to identify how CO4
relationship between size of the accuracy of the model is
the tree and overfitting. affected with increase in
number of nodes.
PART C-SHORT ANSWER QUESTIONS
1 Describe the decision tree Remember —- CO4
learning method?
2 Define hypothesis space for Remember —- CO4
decision tree learning.
3 Explain the inductive bias Understand The learner to recall CO4
considered for decision tree inductive bias and how is it
learning method. incorporated in the learning
process using decision trees.
4 Explain the improvement of Understand The learner to recall the CO5
readability in decision tree different ways in which
learning method. decision trees are
represented for easy
understanding.
5 Mention some applications Remember —- CO12
where decision tree learning
is effectively applied.
6 Explain the measure used to Understand The learner to recall the CO5
identify the best attribute different ways in which best
for classification? attribute at a particular
node is selected.
7 Define entropy and its use Remember —- CO4
in decision tree learning?
8 Explain the significance of Understand The learner to recall the CO4
the hypothesis search ID3 algorithm and the
strategy used in ID3 measures used to find the
algorithm. best attribute in order to
derive the hypothesis.
Page 12
9 Define heuristic search and Remember —- CO4
its role in decision tree
learning?
10 State underfitting in Remember —- CO5
machine learning.
11 State overfitting in machine Remember —- CO5
learning.
12 Describe the purpose of Understand The learner to recall the CO1
validation set in machine different types of datasets
learning? used during developing a
machine learning model.
13 Mention the two classes into Remember —- CO5
which the different
approaches to avoid
overfitting in decision trees
are grouped.
14 Explain the drawback of Understand The learner to relate CO4
reduced error pruning? overfitting to pruning a
decision tree.
15 Describe rule post pruning Understand The learner to state CO4
method in decision tree different methods to avoid
learning? overfitting.
16 Define SplitInformation Remember —- CO4
measure with its formula.
17 Mention the different ways Remember —- CO5
in which missing attribute
values can be handled in
decision tree learning.
18 Recallinductive bias in Remember —- CO5
decision tree learning with
an example.
19 Explain the need of decision Understand The learner to recall the CO4
tree pruning. problem of overfitting and
the ways to overcome it.
20 Distinguish between Analyze The learner to mention the CO1
training set and test set. different types of data set in
terms of size, applicability
and specifications.
Page 13
MODULE III
ARTIFICIAL NEURAL NETWORKS
PART-A PROBLEM SOLVING AND CRITICAL THINKING QUESTIONS
1 Demonstrate a neural Apply The learner to recall ANN CO8
network representation for architecture.
self driving vehicle with a
neat representation of the
architecture.
2 Derive XOR boolean Apply The learner to recall the CO8
function implementation perceptronfunctionality.
using a perceptron with a
neat diagram.
3 Illustrate the visual Apply The learner to represent the CO7
representation of the error surface in terms of
hypothesis space in the weights used...
gradient descent algorithm.
Explain how optimal
hypothesis is obtained from
the diagram.
4 Derive the back propagation Analyze The learner to know the CO8
rule for both hidden and concept in weight update
output unit weights. rules.
5 Explain adding momentum Understand The learner to recall CO8
in back propagation momentum concept and
algorithm and relate it to a relate it to updating
real time example. weights.
CIE II
6 Demonstrate to test a Apply The learner to recall sample CO3
hypotheses h whose error error and true error.
D(h) is known to be in the
range between 0.2 and
0.6.What is the minimum
number of examples you
must collect to assure that
the width of the two-sided
95% confidence interval will
be smaller than 0.1?
Page 14
7 Supposehypotheses h Apply The learner to recall sample CO3
commits r = 10 errors over error, true error and
a sample of n=65 confidence interval for
independently drawn one-sided interval.
examples. What is the 95%
one-sided interval (i.e., what
is the upper bound U such
that error D(h)≤5 U with
95% confidence)?
8 Supposeyou test a Apply The learner to recall basics CO3
hypotheses h and find that of sampling theory.
it commits r=300 errors on
a sample s of n=1000
randomly drawn test
examples. What is the
standard deviation in
errors(h)?
9 A telecom company has Analyze The learner to recall how to CO5
developed a classifier for find out accuracy and
detecting whether a precision based on
cell-phone networkbase hypotheses.
station is faulty or not. (a)
Evaluate the classifier on a
test set. Here is the
confusion matrix. (In the
table, Fmeans faulty and
NF not faulty.) Compute
the accuracy of the
classifier.
Page 15
10 Suppose hypotheses h Apply The learner to recall sample CO3
commits r = 12 errors over error, true error and
a sample of n=67 confidence interval for
independently drawn two-sided interval.
examples. i) What is the
variance and standard
deviation for number of true
error rate errorD(h). ii)
What is the 90% confidence
interval (two-sided) for
thetrue error rate?
PART-B LONG ANSWER QUESTIONS
1 Describe the significance of Understand The learner to recall the CO8
Artificial Neural Network in concept ANN and how it is
Machine Learning. used to implement machine
learning models.
2 Explain the representational Understand The learner should know CO6
power of the perceptron. the linearly separable and
unseparable datasets.
3 Derive AND boolean Apply The learner to recall the CO6
function implementation perceptron functionality.
using a perceptron.
4 Derive OR boolean function Apply The learner to recall the CO6
implementation using a perceptron functionality.
perceptron.
5 Describe perceptron Understand The learner to know the CO6
learning rule of solving a various learning mechanisms
learning problem. available.
6 Explain gradient descent Understand The learner to know the CO7
rule of solving a learning various learning mechanisms
problem. available.
7 Derive NOT boolean Apply The learner to recall the CO6
function implementation perceptron functionality.
using a perceptron.
8 Explain the drawback of Understand The learner to define CO6
perceptron and the way to alternative building units of
overcome it. ANN.
9 Mention the characteristics Remember —- CO8
of the problems appropriate
for ANN representation.
10 Describe the boolean Remember —- CO5
function in disjunctive
formal form with a suitable
example.
Page 16
CIE II
11 What are the steps involved Remember —- CO4
in hypotheses testing?
12 Write a Procedure to Remember —- CO4
estimate the difference in
error between two learning
methods LA and LB.
13 Suppose a computer Apply The learner to recall the CO4
program for recognizing values of confidence matrix
dogs in photographs and find TP, FP and FN to
identifies eight dogs in a find precision and recall.
picture containing 12 dogs
and some cats. Of the eight
dogs identified, five actually
are dogs while the rest are
cats. Compute the precision
and recall of the computer
program.
14 Find 95% confidence Apply This would require the CO3
interval for hypothesis h learner to recall the values
that observed to have 10 of confidence interval.
miss-classifications over a
sample of 100 instances.
15 Assume the following: A Apply The learner to recall the CO4
database contains 80 values of confidence matrix
records on a particular topic and find the required scores
of which 55 are relevant to a of search.
certain investigation. A
search was conducted on
that topic and 50 records
were retrieved. Of the 50
records retrieved, 40 were
relevant. Construct the
confusion matrix for the
search and calculate 1.
Precision 2. Recall 3.
Accuracy 4. Error rate 5.
Sensitivity 6. Specificity 7.
F-measure scores for the
search.
16 Define confusion matrix Remember —- CO4
with suitable example
Page 17
17 How can we interpret the Understand The learner to recall how CO4
”learning of a concept” as learned concept using a
”searching a concept”? learning algorithm is used
for searching a concept in
hypotheses.
18 Explain Binomial Remember —- CO3
Distribution with an
example.
19 Explain most general Understand The learner relates the CO4
hypotheses and most hypotheses related to most
specific hypotheses in the general or specific
context of a classification hypotheses depending on
problem. requirement.
20 Explain the terms i) Type I Understand The learner to differentiate CO4
and Type II errors ii) Null the terms like Type I and II
Hypotheses and Alternate errors and Null and
hypotheses Alternate hypotheses.
PART C-SHORT ANSWER QUESTIONS
1 Explain Sigmoid function Understand The learner to recall the CO6
and its role in the ANN. effect of sinusoidal
activation function.
2 Describe the structure of Understand The learner to represent the CO8
Artificial Neural Network. units and their
interconnection among
them.
3 List out some of the Remember —- CO8
applications of ANN.
4 Define Perceptron. State Remember —- CO6
the output function.
5 Explain the way Understand The learner to recall the CO6
thresholding is implemented effect of step activation
in a perceptron. function.
6 Describe perceptron in Understand The learner to mention the CO6
terms of vectors. vector representation form
of output function.
7 Explain the space of Understand The learner to define the CO4
candidate hypotheses hypothesis obtained.
considered in perceptron
learning.
8 Describe the significance of Understand The learner to know CO1
hyperplanes in machine classification of data points.
learning.
9 Mention different methods Remember —- CO6
of perceptron learning.
Page 18
10 Mention the steps in Remember —- CO7
gradient descent algorithm
for training linear units.
CIE II
11 Write short notes on Understand The learner to recall the CO4
estimating hypotheses concepts of their role in
accuracy obtaining hypothesis
accuracy.
12 What do you mean by Understand Relate constraints on CO4
conjunction and disjunction training data obtains
of constraints while different hypotheses
evaluating hypotheses?
13 What is Central Limit Remember —- CO3
Theorem?
14 What does it mean by a Analyze The learner to recall the CO4
hypothesis that is consistent concepts to make
with training data versus hypotheses consistent
hypothesis that satisfies a
certain instance?
15 Explain the two key Remember —- CO4
difficulties that arise while
estimating the Accuracy of
Hypothesis.
16 Define hypotheses space Understand Relate how the hypotheses CO4
space obtained used for
predicting.
17 What is binomial Remember —- CO3
distribution
18 Define the following terms 1. Understand The learner to know the CO7
Sample error 2. True error difference between the
errors raised on sample and
at the time of prediction.
19 What is confusion matrix? Remember —- CO5
20 Define the following terms i) Remember —- CO3
Mean ii) Variance iii)
Standard Deviation
MODULE IV
BAYESIAN LEARNING
PART A-PROBLEM SOLVING AND CRITICAL THINKING QUESTIONS
1 Give a distribution for P(h) Understand The learner to recall about CO9
and P(D/h) under which MAP hypothesis and find G.
Find G is guaranteed to
output a MAP hypothesis.
Page 19
2 Consider a medical Apply The learner to recall about CO9
diagnosis problem in which MAP hypothesis and
there are two alternative determine the value for
hypotheses: 1. That the given example.
patient has a particular
form of cancer (+) and 2.
That the patient does not
(-). A patient takes a lab
test and the result comes
back positive. 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. Furthermore, .008
of the entire population
have this cancer. Determine
whether the patient has
Cancer or not using MAP
hypothesis.
3 Give a distribution for P(h) Understand The learner to recall about CO9
and P(D/h) under which MAP hypothesis.
Find G is not guaranteed to
output a MAP hypothesis.
4 How do we calculate Understand The learner to recall about CO9
frequency and likelihood text using naı̈ve bayes
tables for a given dataset in theorem.
the “Naı̈ve Bayes”
algorithm
5 Write down the expression Understand The learner to recall about CO9
for the quantity to be text usingminimum
minimized according to the description length principle.
minimum description length
principle.
6 Give a distribution for P(h) Analyze The learner to recall about CO9
and P(D/h) under which MAP find G using ML
Find G is guaranteed to hypothesis.
output a ML hypothesis but
not a MAP hypothesis.
Page 20
7 Is it possible to construct a Understand The learner to recall about CO4
set of training data such text using find G using
that a consistent hypothesis MDL hypothesis.
exists, but MDL chooses a
less consistent hypothesis?
If so, give such a training
set. If not, explain why not.
8 Give probability Analyze The learner to recall about CO4
distributions for P(h) and text using, find G using
P(D/h) such that the above MAP hypothesis.
MDL algorithm outputs
MAP hypotheses.
9 Let XX be a continuous Apply The learner to recall about CO9
random variable with the how to use MAP hypothesis.
following PDF:
Find the MAP estimate of
X given Y=3.
Page 21
5 Explain prior probability Understand The learner torecall CO9
likelihood and marginal probability likelihood and
likelihood in context of marginal likelihood in
Naı̈ve Bayes algorithm? context of Naı̈ve
Bayesalgorithm.
6 ExplainBayesianbelief Understand The learner recalls about CO9
network? Where are they Bayesianbelief network and
used? its uses.
7 Describe about prior Understand The learner torecallprior CO9
probability likelihood in probability likelihood in
NaiveBayes algorithm Naı̈ve Bayes algorithm.
8 How do you classify text Remember —- CO9
using Bayes Theorem?
9 Explain the general cause of Understand The learner to recall about CO5
overfitting and underfitting? overfitting and underfitting.
What steps will you take to
avoid overfitting and
underfitting?
10 Describe the average Understand The learner to recall about CO9
squared difference between classifier predicted output
classifier predicted output and actual output.
and actual output
11 Explain probability of a Understand The learner to recall about CO4
hypothesis before the probability of hypothesis.
presentation of evidence
12 What are the uses of Bayes Remember —- CO9
classifier
13 Explain hypothesis with Understand The learner to relate CO4
example hypothesis for the given
example.
14 What are the advantages of Remember —- CO9
naı̈ve bayes classifier
15 ExplainBayesian belief Understand The learner to recall about CO9
network and conditional Bayesian network and
independence with example. conditional independence.
16 Explain the concept of EM Remember —- CO9
Algorithm.
17 Discuss what are gaussian Understand The learner to recall about CO9
mixtures. gaussian mixtures.
18 What is optimal classifier Remember —- CO9
19 Explain about different Understand The learner to recall about CO9
classifiers different classifiers.
Page 22
20 How do you classify text Understand The learner to recall Bayes CO9
using Bayes theorem theorem and know how to
classify using it.
PART C- SHORT ANSWER QUESTIONS
1 Explain naı̈ve Bayes Remember —- CO9
classifier algorithm?
2 What is the need of Understand The learner to recall about CO1
learning? need of Learning.
3 Define Bayesian Belief nets Understand The learner to recall about CO9
and their uses? Bayesian belief networks.
4 Define (i) Prior Probability Remember —- CO3
(ii) Conditional Probability
(iii) Posterior Probability
5 Explain conditional Understand The learner to recall about CO9
Independence? conditional independence.
6 Describe the concept of Understand The learner to state about CO9
MDL? concept of MDL.
7 Discuss maximum likelihood Understand The learner to recall about CO3
and least square error maximum likelihood and
hypothesis? least square error
hypothesis.
8 Describe maximum Understand The learner to recall about CO3
likelihood hypothesis for maximum likelihood and
predicting probabilities? least square error
hypothesis.
9 Explain naı̈ve Bayes Understand The learner to recall about CO9
classifier with an example? Naı̈ve Bayes classifier and
explain with example.
10 Describe the concept of Remember —- CO9
MDL?
11 Explain brute force Bayes Understand The learner to recall about CO9
concept learning? brute force Bayes concept
learning.
12 Why is naive Bayes so Remember —- CO9
naı̈ve?
13 Define Bayesian theorem? Remember —- CO3
What is the relevance and
features of Bayesian
theorem? Explain the
practical difficulties of
Bayesian theorem.
Page 23
14 Explain the gradient search Understand The learner needs to know CO7
to maximize likelihood in a how to use gradient search
neural network? to maximize likelihood.
15 Explain the concept of EM Understand The learner needs to explain CO9
Algorithm with example? about concept of EM
algorithm.
16 What are the advantages of Remember —- CO9
Naive Bayes?
17 Explain the concept of EM Remember —- CO9
algorithm?
18 Compare conditional Understand
The learner to recall both CO3
probability and joint these algorithms and their
probability? applicability and drawbacks.
19 How does naı̈ve bayes Understand The learner to recall about CO9
classifier works? working functionality of
Naı̈ve Bayes classifier
20 Explain bayes errors? Understand The learner to recall about CO9
Bayes errors and
understand.
MODULE V
INSTANCE BASED AND REINFORMENT LEARNING
PART A-PROBLEM SOLVING AND CRITICAL THINKING QUESTIONS
1 Explain the Q function and Understand The learner to explain CO11
Q learning algorithm deterministic Q learning
assuming deterministic algorithm based on rewards
rewards and actions with and actions.
example.
2 Explain the K – nearest Understand The learner to explain k-nn CO10
neighbor algorithm for classification learning
approximating a discrete – algorithm based on discrete
valued function f : Hn V inputs.
with pseudo code.
3 Compare unsupervised Analyze The learner to recall the CO1
learning and reinforcement different types of learning
learning with examples. and compare with examples.
4 Develop a Q learning task Apply The learner to recall the Q CO12
for recommendation system learning task for online
of an online shopping shopping website.
website. What will be the
environment of the system?
Write the cost function and
value function for the
system.
Page 24
5 Identify the suitable Analyze The learner to recall CO12
learning method for training learning methods and
a robotic arm and explain identify suitable learning
it. method for training a
robotic arm.
6 Consider playing Analyze The learner to recall CO12
Tic-Tac-Toa against an tic-tac-toe game and
opponent who plays identify states, transitions
randomly. In particular, and rewards in MDP.
assume the opponent
chooses with uniform
probability any open space,
unless there is a forced
moved (in which case it
makes the obvious correct
move). (a) Formulate the
problem of learning an
optimalTic-Tac-Toa
strategy in this case as a
Q-learning task. (b) What
are the states, transitions,
and rewards in this
non-deterministic Markov
decision process?
7 How does Q function Understand The learner to relate Q CO11
become able to learn with function learning with and
and without complete without complete
knowledge of reward Knowledge of reward
function and state function and state
transition function. transition function.
8 How does setting a Understand The learner should CO11
Reinforcement Learning understand parameters
problem require an needed for reinforcement
understanding of the learning.
following parameters of the
problem? (a) Delayed
reward (b) Exploring
unknown or exploiting
already learned states and
actions. (c) Number of old
states should be considered
to decide action
Page 25
9 Consider playing Understand The learner to recall CO11
Tic-Tac-Toa against an tic-tac-to a game and
opponent who plays analyze will your game
randomly. In particular, succeed if the opponent
assume the opponent plays optimally rather than
chooses with uniform randomly.
probability any open space,
unless there is a forced
moved (in which case it
makes the obvious correct
move). (a) Outline Will
your program succeed if the
opponent plays optimally
rather than randomly.
10 Identify the suitable Apply The learner to recall CO12
learning method for learning methods and
applications given below identify suitable learning
and explain how the method for the applications.
learning method is used. i)
Videos Surveillance ii)
Predictions while
Commuting iii) Industry
automation
PART B-LONG ANSWER QUESTIONS
1 Explain the Q function and Understand The learner to recall Q CO11
Q Learning Algorithm. function and know how it is
used in Q learning
algorithm.
2 Describe K-nearest Understand The learner to learn using CO10
Neighbor learning k-nn algorithm on the real
Algorithm for continues valued data for classification
(real) valued target problems.
function.
3 What are different aspects Remember —- CO11
of the environment that
points towards
reinforcement learning?
4 Explain Locally Weighted Understand The learner to learn using CO10
Linear Regression. locally weighted linear
regression.
5 Explain about Understand The learner to learn the CO10
distance-weighted nearest refinement to the k-nn
neighbor algorithm. algorithm by giving weights.
Page 26
6 Discuss the major Understand The learner to identify the CO10
drawbacks of K-nearest drawbacks of k-nn
Neighbor learning algorithm.
Algorithm and how it can
be corrected.
7 Explain difference between Understand The learner to memorize the CO1
lazy learning and eager differences between lazy
learning. learning and eager learning.
8 Recall Q function and Analyze The learner to relate Q CO11
algorithm for Q learning, learning to dynamic
how does it relate to programming.
dynamic programming?
9 What is reinforcement Understand The learner todesign the CO11
learning and explain aspects and describe the
reinforcement learning learning problem using RL
problem with neat diagram. method.
10 Explain how an agent can Understand The learner to explain how CO11
take action to move from agent will move from one
one state to other state with step to other with the help
the help of rewards. of rewards.
11 Choose Q learning in Remember —- CO11
deterministic environment...
12 Explain reinforcement Understand The learner to understand CO11
learning in detail along with various elements involved in
the various elements reinforcement learning.
involved in forming the
concept.
13 Explain Convergence of Q Understand The learner to recall CO11
learning for deterministic convergence theorem and
Markov decision process use it in Q learning.
theorem with proof.
14 Explain how reinforcement Understand The learner to relate how CO11
learning problem differs reinforcement learning
from other function differs from other learning.
approximation. convergence theorem and
use it in Q learning.
15 Explain CADET System Understand The learner to recall Case CO11
using Case based reasoning. based reasoning how we can
use in CADET System
16 Interpret the role of radial Understand The learner to recall how we CO11
basis function in can we use radial basis
separatingnonlinear function for nonlinear
patterns. patterns.
Page 27
17 Explain about Q learning in Understand The learner to recall Q CO11
non deterministic learning and to know how
environment. to use in non deterministic
environment.
18 Whatis instance based Remember —- CO1
learning? Explain key
features and disadvantages
of these methods.
19 Explain the k-nearest Understand The learner to recall k-nn CO10
neighbor algorithm for algorithm behavior for
appropriate in a discrete inputs.
discrete-valued function.
20 Explain the behavior of an Understand The learner to understand CO11
agent in a Markov decision agent behavior in markov
process (MDP). decision process.
PART C-SHORT ANSWER QUESTIONS
1 What is instance based Remember —- CO11
learning
2 Explain instance based Understand The learner should identify CO11
learning methods. instance based learning
methods.
3 Explain convergence Understand The learner to recall CO3
theorem of Q learning for convergence theorem and
non deterministic Markov use it for MDP.
decision process (MDP).
4 What is Reinforcement Remember —- CO11
Learning?
5 Analyze aboutK-nearest Analyze The learner should relate CO10
Neighbor learning K-NN algorithm to any one
Algorithm and give reason of category of ML.
under which category of
Machine learning algorithm
it comes?
6 Remarks on Lazy and Eager Understand The learner should CO1
Learning. understand the drawbacks
of lazy and eager learning
methods.
7 Explain properties of Understand The learner to relate CO12
case-based reasoning properties to case-based
system. reasoning system.
8 What is radial basis Remember —- CO12
function.
Page 28
9 Write reinforcement Understand The learner to recall the CO11
learning problem usage reinforcement learning
characteristics. for different problems.
10 Define radial basis function Remember —- CO10
network.
11 What are lazy learning Remember —- CO10
methods?
12 What is temporal difference Remember —- CO1
learning?
13 Explain Convergence in Q Understand The learner to use CO11
learning. convergence theorem in Q
learning.
14 Define the following terms Remember —- CO10
with respect to K - Nearest
Neighbor Learning: 1.
Regression 2. Residual 3.
Kernel Function
15 What is Q function? Remember —- CO11
16 Explain 3 key properties of Understand The learner should CO10
instance based methods understand the properties of
instance based methods.
17 What is Bellman’s equation Remember —- CO11
and why we use it.
18 Describe the elements used Understand The learner to identify the CO11
in reinforcement learning? elements used in
reinforcement learning.
19 Explain about offline Understand The learner should analyze CO12
systems and online systems. the characteristics related to
offline and online systems.
20 Show how partially Understand The learner to know how to CO11
observed data used in use reinforcement learning
reinforcement learning? on partially observed data.
Page 29