Course

Unit I

Introduction to ML; Problems, data and tools. Learning systems, goals, challenges and applications of the machine learning systems. Aspects of developing system, training data, testing data, concept representation, classification errors, validation. Dimensionality Reduction, Data compression, PCA.

Unit II

Linear regression, SSE, gradient descent, bias and variance estimation, overfitting and underfitting, regularization, ridge and lasso regression.

Unit III

Logistic regression, hypothesis representation, decision boundary, cost function, multi-class classification. Nearest neighbour methods. Decision Tree learning, representing concepts as decision trees, picking the best splitting attribute: entropy and information gain. Probability and classification, Naïve Bayes classification, EM algorithm, kernels, Kernel regression, kernels, Support vector machine (SVM) and kernels, kernel optimization. Linear Discriminant Analysis algorithm.

Unit IV

Neural networks learning, non-linear hypothesis, model representation, perceptron, cost function, back propagation algorithm.

Unit V

Unsupervised learning, clustering, different clustering methodologies. Current problems on Machine Learning.

This course will enable students to understand the basic concepts of machine learning. It will help students to apply different machine learning models to real-world problems.

Course Objectives

• To understand basic concepts of machine learning
• To familiarize the machine learning models like linear and logistic regression
• To understand different classifiers
• To understand different clustering algorithms

Course Outcomes

1. Machine Learning, Tom Mitchell, McGraw Hill, 1997.
2. Duda, Richard, Peter Hart, and David Stork, “Pattern Classification” Second Edition, New York, NY: Wiley-Interscience, 2000.
3. Hastie, T., R. Tibshirani, and J. H. Friedman, “The Elements of Statistical Learning: DataMining, Inference and Prediction”, New York, Springer, 2001
4. Christopher, M. Bishop. Pattern Recognition and Machine Learning, Springer-Verlag New York, 2016.