JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD

Prerequisites

• Students are expected to have knowledge basic linear algebra, basic probability theory and basic programming techniques;
• A course on “Computational Mathematics”
• A course on “Computer Oriented Statistical Methods”

Course Objectives

• This course introduces fundamental concepts, theories, and algorithms for pattern recognition and machine learning.
• Topics include: Pattern Representation, Nearest Neighbor Based Classifier, Bayes Classifier, Hidden Markov Models, Decision Trees, Support Vector Machines, Clustering, and an application of hand-written digit recognition.

Course Outcomes

• Understand the theory, benefits, inadequacies and possible applications of various machine learning and pattern recognition algorithms
• Identify and employ suitable machine learning techniques in classification, pattern recognition, clustering and decision problems.

UNIT - I: Introduction

What is Pattern Recognition, Data Sets for Pattern Recognition, Different

Paradigms for Pattern Recognition. Representation: Data Structures for Pattern Representation, Representation of Clusters, Proximity Measures, Size of Patterns, Abstractions of the Data Set, Feature Extraction, Feature Selection, Evaluation of Classifier, Evaluation of Clustering.

UNIT - II: Nearest Neighbor Based Classifier

Nearest Neighbor Algorithm, Variants of the NN Algorithm

use of the Nearest Neighbor Algorithm for Transaction Databases, Efficient Algorithms, Data Reduction, Prototype Selection. Bayes Classifier: Bayes Theorem, Minimum Error Rate Classifier, Estimation of Probabilities, Comparison with the NNC, Naïve Bayes Classifier, Bayesian Belief Network.

UNIT - III: Hidden Markov Models

Markov Models for Classification, Hidden Morkov Models,

Classification using HMMs. Decision Trees: Introduction, Decision Tree for Pattern Classification, Construction of Decision Trees, Splitting at the Nodes, Overfitting and Pruning, Examples of Decision Tree Induction.

UNIT - IV: Support Vector Machines

Introduction, Learning the Linear Discriminant Functions, Neural

Networks, SVM for Classification. Combination of Classifiers: Introduction, Methods for Constructing Ensembles of Classifiers, Methods for Combining Classifiers.

UNIT - V: Clustering

Why is Clustering Important, Hierarchical Algorithms, Partitional Clustering,

Clustering Large Data Sets. An Application-Hand Written Digit Recognition: Description of the Digit Data, Preprocessing of Data, Classification Algorithms, Selection of Representative Patterns, Results.

TEXT BOOK:

• 1. Pattern Recognition: An Algorithmic Approach: Murty, M. Narasimha, Devi, V. Susheela, Spinger Pub,1st Ed.

REFERENCES:

• 1. Machine Learning - Mc Graw Hill, Tom M. Mitchell.
• 2. Fundamentals Of Speech Recognition: Lawrence Rabiner and Biing- Hwang Juang. Prentice- Hall Pub. R18 B.TECH IT III YEAR