Pattern Recognition (PEC-IT602D)
9 units: Bayesian decision theory, Parameter estimation methods, Hidden Markov models for sequential pattern classification, Dimension reduction methods, Linear discriminant function based classifiers · L-T-P 3L/week.
Detailed syllabus
Unit 1 (2 hours)
Basics of pattern recognition
Unit 2: Bayesian decision theory (8 hours)
Bayesian decision theory: classifiers, discriminant functions, decision surfaces, normal density and discriminant functions, discrete features
Unit 3: Parameter estimation methods (6 hours)
Parameter estimation methods: maximum-likelihood estimation, Gaussian mixture models, expectation-maximization method, Bayesian estimation
Unit 4: Hidden Markov models for sequential pattern classification (8 hours)
Hidden Markov models for sequential pattern classification: discrete hidden Markov models, continuous density hidden Markov models
Unit 5: Dimension reduction methods (3 hours)
Dimension reduction methods: Fisher discriminant analysis, principal component analysis; Parzen-window method, k-nearest neighbour method
Unit 6 (2 hours)
Non-parametric techniques for density estimation
Unit 7: Linear discriminant function based classifiers (5 hours)
Linear discriminant function based classifiers: perceptron, support vector machines
Unit 8: Non-metric methods for pattern classification (4 hours)
Non-metric methods for pattern classification: non-numeric (nominal) data, decision trees
Unit 9: Unsupervised learning and clustering (2 hours)
Unsupervised learning and clustering: criterion functions for clustering, k-means, hierarchical and other clustering algorithms
Source: official document