Machine Learning (PCCST503) Syllabus

Introduction to ML: machine learning vs. traditional programming, machine learning paradigms — supervised, semi-supervised, unsupervised, reinforcement learning. Parameter Estimation: maximum likelihood estimation (MLE) and maximum a-posteriori estimation (MAP), Bayesian formulation.

Supervised Learning: feature representation and problem formulation, role of loss functions and optimization. Regression: linear regression with one variable, linear regression with multiple variables — solution using gradient descent and matrix method. Classification: logistic regression, Naive Bayes, KNN, decision trees (ID3). Generalisation and Overfitting: idea of overfitting, LASSO and RIDGE regularization, training/testing/validation. Evaluation measures: classification (precision, recall, accuracy, F-measure, ROC, AUC) and regression (MAE, RMSE, R-squared).

SVM: linear SVM, idea of hyperplane, maximum margin hyperplane, non-linear SVM, kernels for learning non-linear functions. Neural Networks (NN): perceptron, multilayer feed-forward network, activation functions (Sigmoid, ReLU, Tanh), backpropagation algorithm.

Unsupervised Learning: clustering — similarity measures, hierarchical clustering (agglomerative), partitional clustering, K-means clustering. Dimensionality reduction: Principal Component Analysis, multidimensional scaling. Ensemble methods: bagging, boosting. Resampling methods: bootstrapping, cross validation. Practical aspects: bias-variance tradeoff.