Course Details
Subject {L-T-P / C} : EE4601 : Introduction to Machine Learning { 3-0-0 / 3}
Subject Nature : Theory
Coordinator : Dr. Shekha Rai
Syllabus
• Module 1: Introduction (1 hour)
o Definition
o Examples
o Process of ML
o Applications of ML
• Module 2: Types of learning and Evaluation (3 hours)
o Supervised
o Unsupervised
o Reinforcement
o Inductive learning and hypothesis space
o Evaluation
o Cross-validation
• Module 3: Regression (3 hours)
o Linear algebra and least square
o Types of regression models
• Module 4: Decision tree (3 hours)
o Entropy
o Information gain
o Building a decision tree
o Overfitting-reduce by prepruning, post-pruning
• Module 5: kNN (2 hour)
o Defination
o Use of kNN in data –set
o Advantages
o Disadvantages
• Module 6: Feature reduction (3 hours)
o PCA
o LDA
o Advantage of LDA over PCA
• Module 7: Bayesian learning (5 hours)
o Review of probability theory
o Bayesian concept learning
o Bayes optimal classifier
o Naive Bayes classifier
o Bayesian Belief Network
• Module 8: Support vector machines (8 hours)
o Classification using hyperplanes
o Identifying the correct hyperplane in SVM
o Maximum margin hyperplane
o Functional margin
o Geometric margin
o Maximize margin width
o Large margin linear classifier
o Dual formulation
o Maximum Margin with noise
o Nonlinear SVM and Kernel function
o Multi-class classification
o Sequential minimal optimization
• Module 8: Neural network (5 hours)
o Artificial neural network
o Perceptron training rule
o Gradient descent
o Multilayer neural network
o Backpropagation training algorithm
o Deep neural network
o Convolutional neural network
• Module 9: Clustering (2 hours)
o Partitional clustering
o Hiearchial clustering
o Mixture of Gaussian
• Module 10: Computational learning theory (4 hours)
o Prototypical concept learning task
o PAC model
o Sample complexity of supervised learning
o Sample complexity: inconsistent finite hypothesis
o S algorithm
o Infinite hypothesis space
o VC dimension
• Module 11: Ensemble learning (1 hour)
o Definition
o Its challenges
Course Objectives
- To be able to formulate machine learning problems corresponding to different applications.
- To understand a range of machine learning algorithms along with their strengths and weaknesses.
- To understand the basic theory underlying machine learning
- To be able to apply machine learning algorithms to solve problems of moderate complexity.
Course Outcomes
At the end of the course, students will be able to <br />1: Understand the fundamental issues and challenges of machine learning: data, model selection, model complexity, etc. <br />2: Differentiate between supervised and unsupervised learning. <br />3: Understand the strengths and weaknesses of many popular machine learning approaches. <br />4: Appreciate the underlying mathematical relationships within and across Machine Learning algorithms. <br />5: Choose a suitable machine learning algorithm for real world applications
Essential Reading
- Tom M Mitchell, Machine Learning, PHI , 2015
- Ethem Alpaydin, Introduction to Machine Learning, The MIT Press , 3rd Edition 2015
Supplementary Reading
- Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani, Introduction to Statistical Learning, Springer, 2013
- Richard Duda, Peter Hart, David Stork, Pattern Classification, John Willey
Journal and Conferences
- IEEE Transaction on Pattern Analysis and Machine Intelligence
- IEEE Transaction on Neural Networks & Learning Systems