Statistical Pattern Recognition and Decision Making Methods

Course guarantor:

Lecturer:

Tutor:

Supervisor:

Department of Software Engineering

Synopsis:

Collection of recognition and classification methods with accent to mathematical and statistical principles of their design

and functionality.

Requirements:

Syllabus of lectures:

1.Introduction - what is pattern recognition and decision making

2.Statistical (feature-based) and structural (syntactic) pattern recognition

3.Introduction to statistical pattern recognition - supervised and non-supervised classifiers

4.Simple metric classifiers - NN classifier, k-NN classifier, linear classifier

5.Bayesian classifier - the basic principle, parametric and non-parametric B.c., B.c. for normally distributed classes, parameter estimation, necessary conditions of linearity, special cases in two dimensions

6.Non-metric classifiers, decision trees

7.Non-supervised classifiers - cluster analysis in the feature space, iterative and hierarchical methods, criteria of cluster separability

8.k-means iterative algorithm and its modifications

9.Agglomerative hierarchical clustering, inter-cluster metrics, stop conditions, estimating the number of clusters

10.Dimensionality reduction of the feature space, feature extraction and selection, class separability criteria, Mahalanobis distance

11.Principal component transform

12.Optimal and sub-optimal feature selection methods, sequential and floating search

13.Decision making as a discrete optimization problem

14.Basic methods for unconstrained and constrained discrete optimization

Syllabus of tutorials:

Study Objective:

Study materials:

Key references:

[1] Urbanowicz, R. J. J., Browne, W. N. Introduction to Learning Classifier Systems. Berlin: Springer, 2017.

[2] Matloff, N. Statistical Regression and Classification: From Linear Models to Machine Learning. Boca Raton: CRC

press, 2017.

Recommended references:

[3] Duda, R. O., Hart, P. E., Stork, D. G. Pattern Classification. 2nd edition. New York: Willey, 2007.

[4] Scholkopf, B., Smola, A. J. Learning with Kernels. Cambridge: MIT Press, 2001.

[5] Aggarwal, Ch. C. Data Mining: The Textbook. Cham (Switzerland): Springer, 2015.

[6] Izenman, A. J. Modern Multivariate Statistical Techniques: Regression, Classification, and Manifold Learning.

Corr. 2nd printing 2013 edition. New York: Springer, 2013.

[7] Proceedings of the International Workshop on Multiple Classifier Systems (MCS).

Note:

Further information:

No time-table has been prepared for this course

The course is a part of the following study plans:

• Aplikace informatiky v přírodních vědách (compulsory course in the program)