Pattern Recognition

The objective of the course is to introduce the fundamentals required for understanding tools for practical implementation and metrics for performance analysis of classical pattern recognition algorithms.

Upon completion of the course, the student should be able to

• Formulate a supervised and unsupervised learning problem (Unit 1)

• A student should be able to select and implement a suitable pattern recognition algorithm for simulated data (Units 2,3)

• Design and code suitable feature engineering methods

• To implement and evaluate the performance of classification algorithms on real-world datasets

1. Introduction (5 hrs)

   1. Introduction: Review of Probability theory, Probability distributions, Decision theory, evaluation criteria, Supervised and Unsupervised Learning

2. Linear Models (5 hrs)

   1. Discriminant functions, Generative models, Probabilistic Discriminative Models, Maximum likelihood approach, Likelihood ratio test (2 Weeks)

3. Parametric Classifiers (10 hrs)

   1. SVM Theory and Implementation: Theory of SVMs, Binary classification using SVM, Soft-Margin SVM, Kernels in SVMs, Mercer’s theorem, M-class Classification, Binarization, One class SVM

   2. GMMs, EM algorithm, K-means clustering

4. Feature extraction (6 hrs)

   1. Time domain features, Frequency domain features, short time-frequency domain features, Spatial features

5. Non parametric classifiers (5 hrs)

   1. Decision Trees, K Nearest Neighbor classifier, Random Forests

6. Introduction to ANNs for classification (5 hrs )

   1. ANN architectures, Feedforward neural networks, Backpropagation algorithm

Text-Books

1. Pattern Recognition and Machine Learning by C. Bishop (Unit -1 & 3)

2. Pattern Recognition by Sergios Theodoridis (Unit - 2)

3. An Introduction to Support Vector Machines and Other Kernel-based Learning Methods by Nello Christianni (Unit - 3)

4. Feature extraction foundations edited by Isabella Guyon (Unit - 4)

5. E. Alpaydin, Introduction to machine learning, MIT press. (Unit - 5)

6. Neural networks for pattern recognition by C. Bishop. (Unit - 6)

Reference Books

1. Duda, Richard O. and Peter E. Hart. Pattern classification. John Wiley & Sons, 2006.

2. Theodoridis, Sergios, et al. Introduction to pattern recognition: a matlab approach. Academic Press, 2010.

3. Bisong, Ekaba. Building machine learning and deep learning models on Google cloud platform. Berkeley: Apress, 2019.

Computer Assignments

Quizzes

Project

In-class examinations