Pattern Recognition I

Basic concepts of pattern recognition. Supervised and unsupervised training. Estimation of the probability of classification error-Error bounds. Distance functions. Minimum distance pattern classification. knearest neighbour classification. Single and multiply prototypes. Decision functions. Linear decision functions. Perceptron and kmeans algorithm. Bayes classifier. Bayes decision rule for minimum risk. Estimation of probability density function: Maximum entropy criterion, Parzen estimate, orthonormal functions approximation. Stochastic approximation of the probability density function: Robbins-Monro and LMS algorithm. Neural networks structure. Error correction, competitive and hebbian learning. Multilayer perceptron. Back-propagation of error. Radial-Basis function networks.Hopfield machine. Syntactic pattern recognition. Formal languages. Type-0,1,2,3. CYK algorithm. Stochastic languages. Grammatical inference. Error correction.

Objectives

The student upon successful completion of the course, expected to know:

Prerequisites

1. Probability & Statistics 2. Matlab

Syllabus

The contents of the course Pattern Recognition I are: Methods of pattern recognition Building Systems Stochastic Systems Neural Networks in Pattern Classification