I M. Tech. - I Sem. (CSE) L T C 3 0 3 Program Elective I (16CS5010) Course Objectives
I M. Tech. - I Sem. (CSE) L T C 3 0 3 Program Elective I (16CS5010) Course Objectives
(CSE) L T C
303
Program Elective I
(16CS5010) MACHINE LEARNING
Course Objectives:
To learn the concept of how to learn patterns and concepts from data without being explicitly programmed in
various IOT nodes.
To design and analyze various machine learning algorithms and techniques with a modern
outlook focusing on recent advances.
Explore supervised and unsupervised learning paradigms of machine learning.
To explore Deep learning technique and various feature extraction strategies.
Course Outcomes:
After completion of course, students would be able to:
Extract features that can be used for a particular machine learning approach in various IOT
applications.
To compare and contrast pros and cons of various machine learning techniques and to get an
UNIT I
Supervised Learning (Regression/Classification)
Basic methods: Distance-based methods, Nearest-Neighbours, Decision
Trees, Naive Bayes Linear models: Linear Regression, Logistic Regression, Generalized Linear Models
Support Vector Machines, Nonlinearity and Kernel Methods
Beyond Binary Classification: Multi-class/Structured Outputs, Ranking
UNIT II
Unsupervised Learning
Clustering: K-means/Kernel K-means
Dimensionality Reduction: PCA and kernel PCA
Matrix Factorization and Matrix Completion
Generative Models (mixture models and latent factor models)
UNIT III
Evaluating Machine Learning algorithms and Model Selection, Introduction to Statistical Learning Theory,
Ensemble Methods (Boosting, Bagging, Random Forests)
UNIT IV
Sparse Modeling and Estimation, Modeling Sequence/Time-Series Data, Deep Learning and Feature Representation
Learning Scalable Machine Learning (Online and Distributed Learning) A selection from some other advanced
topics, e.g., Semi-supervised Learning, Active Learning, Reinforcement Learning, Inference in Graphical Models,
Introduction to Bayesian Learning and Inference.
UNIT V
Recent trends in various learning techniques of machine learning and classification methods for IOT applications,
Various models for IOT applications. R18 M.Tech – CSE
Text /References:
1. Kevin Murphy, Machine Learning: A Probabilistic Perspective, MIT Press, 2012
2. Trevor Hastie, Robert Tibshirani, Jerome Friedman, The Elements of Statistical Learning,
Springer 2009 (freely available online)
COURSE DESCRIPTION
Machine Learning is concerned with computer programs that automatically improve their performance through
experience. This course covers the theory and practical algorithms for machine learning from a variety of
perspectives. We cover topics such as Bayesian networks, decision tree learning, statistical learning methods,
unsupervised learning and reinforcement learning. The course covers theoretical concepts such as inductive bias,
Bayesian learning methods. Short programming assignments include hands-on experiments with various learning
algorithms. This course is designed to give a graduate-level student a thorough grounding in the methodologies,
technologies, mathematics and algorithms currently needed by people who do research in machine learning.
COURSE OBJECTIVES
This course will enable students to,
1. Define machine learning and understand the basic theory underlying machine learning.
2. Differentiate supervised, unsupervised and reinforcement learning
3. Understand the basic concepts of learning and decision trees.
4. Understand neural networks and Bayesian techniques for problems appear in machine learning
5. Understand the instant based learning and reinforced learning
6. Perform statistical analysis of machine learning techniques.
COURSE OUTCOMES
After studying this course, the students will be able to
You're standing at a crowded railway station waiting for a friend. As you wait, hundreds of people
pass you by. Each one looks different, but when your friend arrives you have no problem picking
her out of the crowd. Recognizing people's faces is something we humans do effortlessly, but
how would you program a computer to recognize a person? You could try to make a set of rules.
For example, your friend has long black hair and brown eyes, but that could describe literally
billions of people. What is it about her that you actually recognize? Something about the shape of
her nose or the curve of her chin? But can you put it into words? The truth is that we can
recognize people without ever really knowing how we do it. We cannot describe every detail of
how we recognize someone. We just know how to do it. The trouble is that to program a
computer, we need to break the task down into its little details. That makes it very difficult or even
impossible to program a computer to recognize faces. Face recognition is an example of a task
that people find very easy, but that is very very hard for computers. These tasks are often called
artificial intelligence or AI. You're now going to learn about the type of computing technique that
can solve a lot of these AI problems and that's revolutionising what computers can do that
technique is machine learning.
1 + T T
2 + T T
3 - T F
4 + F F
5 - F T
6 - F T
o (a) What is the entropy of this collection of training examples with respect to the target function
classification?
(b) What is the information gain of a2 relative to these training examples?.
3. NASA wants to be able to discriminate between Martians (M) and Humans (H) based on the following
characteristics: Green ∈{N, Y} , Legs ∈{2,3} , Height ∈{S, T}, Smelly ∈{N, Y}
o Our available training data is as follows:
a) Greedily learn a decision tree using the ID3 algorithm and draw the tree .
b)
(i) Write the learned concept for Martian as a set of conjunctive rules (e.g., if (green=Y and legs=2 and height=T and
smelly=N), then Martian; else if ... then Martian;...; else Human).
(ii) The solution of part b)i) above uses up to 4 attributes in each conjunction. Find a set of conjunctive rules using
only 2 attributes per conjunction that still results in zero error in the training set. Can this simpler hypothesis be
represented by a decision tree of depth 2? Justify.
6. Describe hypothesis Space search in ID3 and contrast it with Candidate-Elimination algorithm.
8. Illustrate Occam’s razor and relate the importance of Occam’s razor with respect to ID3 algorithm.
9. List the issues in Decision Tree Learning. Interpret the algorithm with respect to Overfitting the data.
10. Discuss the effect of reduced Error pruning in decision tree algorithm.
11. What type of problems are best suited for decision tree learning
Machine Learning Module-3 Questions.
2) What are the type of problems in which Artificial Neural Network can be applied.
5) Under what conditions the perceptron rule fails and it becomes necessary to apply the delta rule
12)Derive the Backpropagation rule considering the training rule for Output Unit weights and Training Rule for
Hidden Unit weights
14) Explain how to learn Multilayer Networks using Gradient Descent Algorithm.
15)What is Squashing Function?
4) Explain Brute force MAP hypothesis learner? What is minimum description length principle
7) Define (i) Prior Probability (ii) Conditional Probability (iii) Posterior Probability
3. Describe K-nearest Neighbour learning Algorithm for continues valued target function.
4. Discuss the major drawbacks of K-nearest Neighbour learning Algorithm and how it can be corrected
7.Explain the K – nearest neighbour algorithm for approximating a discrete – valued functionf : Hn→ V with pseudo
code