Machi

ne
Learni
ng
© Oxford University Press 2021. All rights reserved
Introduction to Machine Learning

© Oxford University Press 2021. All rights reserved

Evolution in Business Organizations
 Though with huge data full potential of data was not
utilized because

Data was scattered and able to integrate

Lack of awareness about S/W Tool to extract useful
info from data
 Popularity for Machine Learning
• High volume of data
• Reduced cost of storage
• Availability of complex algorithms and tools
• Need for Machine Learning?
• Need to analyze data for taking decisions
Knowledge Pyramid
Maturity of mind

Keep more stock during festival

Historic patterns and trends

(sales during festival)

Fast selling product

Sales data
Objective of Machine learning
 Take better decisions to design new products
 Improve the business processes
 Develop effective decision support systems
 What is Machine Learning?
“Machine learning allows machines to learn from
data and improve overtime without being explicitly
programmed”

Machine Learning coined in 1959 by ARTHUR

SAMUEL (IBM in computer gaming and AI)
Problem with Conventional Programming
• Conventional programming difficult for many real
world problems
• Ex. Puzzles, games and complex image recognization
 Expert System(AI)
Understand the problems  Develop general purpose rules manually  Formulate
into logic  Implement in a pgm to create a intelligent system
Ex. MYCIN Medical diagnosis (expert knowledge of many doctors into system)

 The above approach was impractical in many domains as programs still

depended on human expertise
 Did not truly exhibit intelligence
 “The momentum shifted to machine learning in the form of data driven systems”
A Learning System for Humans and Machine

The focus of Al is to develop intelligent systems by using data-

driven approach, where data is used as an input to develop
intelligent models. The models can then be used to predict
new inputs.
 What is a Model?
Learning : The process of acquiring knowledge and expertise
through study, experience or being taught

Learning Program summarizes the raw data into a model

A Model is an explicit description of pattern within data in the form of

• Mathematical equation (Linear regression)
• Relational diagrams like graphs/trees
• Logical if/else rules
• Groupings called clusters
 Another View of Machine Learning
• TOM MITCHELL DEFINITION OF MACHINE LEARNING

• "A computer program is said to learn from experience E, with respect to task T and some
performance measure P, if its performance on T measured by P improves with experience E."
• Experience E, Task T, Performance measure P.”
• Ex The task T - Detecting an object in an image.
 Machine Gain the knowledge of object using training dataset of thousands of images.
This is called experience E.
 The ability of the system to detect the object is measured by performance measures P like
precision and recall.
ARTHUR SAMUEL TOM MICHELL
Human Experience
Humans gain experience by various means
• They gain knowledge by rote learning (memorization tech.)
• They observe others and imitate it.
• Humans gain a lot of knowledge from books.
• We learn many things by trial and error.
Once the knowledge is gained, when a new problem is encountered, humans search
for similar past situations and then formulate the heuristics and use that for
prediction.
 Concept: An idea
Models of computer systems that’s formed by
combining all
Computer systems, experience is gathered by these steps:
features or
Collection of data attributes of
something
1. Abstract concepts are formed out of that data.
2 Abstraction is used to generate concepts.
Example, we have some idea about how an elephant looks like.
3. Generalization converts the abstraction into an actionable form of
intelligence.
4. Evaluation checks the thoroughness of the models and to-do course
correction, if necessary, to generate better formulations.
 Machine Learning and AI
• Relation between machine learning and ai • Machine learning is an important branch of Al.
• The aim of Al is to develop intelligent agents.
• An agent can be a robot, humans, or any
autonomous systems
• The focus was on logic and logical inferences.
• It had seen many ups and downs. Al winters.
• Resurgence in Al happened due to development
of data driven systems
• The aim is to find relations and regularities
present in the data.
• Machine learning is the subbranch of Al, whose
aim is to extract the patterns for prediction.
• Deep learning is a subbranch of machine
learning. In deep learning, the models are
constructed using neural network technology
 Machine Learning and Data Science
• Data science is an “umbrella term” covering from data Collection to data analysis.
• Data science and machine learning are interlinked
• Machine learning is a branch of data science.
 Machine Learning and Data Science
• CHARACTERISTICS OF BIG DATA

© Oxford University Press 2021. All rights reserved

 Machine Learning and Data Science
• DATA SCIENCE AND DATA MINING

© Oxford University Press 2021. All rights reserved

 Machine Learning and Data Science
• DATA SCIENCE AND DATA ANALYTICS / PATTERN RECOGNITION

© Oxford University Press 2021. All rights reserved

 Machine Learning and Statistics
ROLE OF STATISTICS

© Oxford University Press 2021. All rights reserved

Machine Learning Types

© Oxford University Press 2021. All rights reserved

Labelled Data

© Oxford University Press 2021. All rights reserved

 Unlabelled Data
DATA THAT IS NOT ASSOCIATED WITH LABELS IS CALLED UNLABELLED DATA.

© Oxford University Press 2021. All rights reserved

 Supervised Learning
CLASSSIFICATION

© Oxford University Press 2021. All rights reserved

 Supervised Learning
KEY ALGORITHMS

© Oxford University Press 2021. All rights reserved

 Supervised Learning
REGRESSION ALGORITHMS

© Oxford University Press 2021. All rights reserved

 Unsupervised Learning
CLUSTERING IS A GROUPING PROCESS.

© Oxford University Press 2021. All rights reserved

 Unsupervised Learning
KEY ALGORITHMS OF UNSUPERVISED LEARNING

© Oxford University Press 2021. All rights reserved

Key Differences

© Oxford University Press 2021. All rights reserved

Semi-supervised Learning

© Oxford University Press 2021. All rights reserved

Reinforcement Learning

© Oxford University Press 2021. All rights reserved

 Challenges of Machine Learning
1. ILL-POSED PROBLEMS – PROBLEMS WHOSE SPECIFICATIONS ARE NOT CLEAR
2. HUGE DATA
3. HUGE COMPUTATION POWER
4. COMPLEXITY OF ALGORITHMS
5. BIAS-VARIANCE

© Oxford University Press 2021. All rights reserved

 Machine Learning Process
MACHINE LEARNING/DATA MINING PROCESS

© Oxford University Press 2021. All rights reserved

 Machine Learning Applications
MACHINE LEARNING MAJOR APPLICATIONS

© Oxford University Press 2021. All rights reserved

 Machine Learning Applications
MACHINE LEARNING MAJOR APPLICATIONS

© Oxford University Press 2021. All rights reserved

Summary

© Oxford University Press 2021. All rights reserved

Summary

© Oxford University Press 2021. All rights reserved