EXAM REQUIREMENTS

The matters discussed during the lectures and exercises are required in the exam. In addition to the distributed course material (lecture slides and exercise answers), the subjects can be studied from the following sources:

INTRODUCTION:

  • Schalkoff, chapter 1 (introduction)
  • Theodoridis and Koutroumbas,
    • chapter 1 (introduction);
    • chapter 5 (feature selection), sections 5.1-5.2.3, 5.4-5.5 (not section on Chernoff Bouns and Bhattarcharyya Distance), 5.6-5.6.2
    • chapter 10 (evaluation), sections 10.1-10.3

STATISTICAL PATTERN RECOGNITION:

  • Schalkoff, chapter 2
  • Theodoridis and Koutroumbas, chapter 2, sections 2.1-2.4

ESTIMATION OF PROBABILITY DISTRIBUTIONS:

  • Schalkoff, chapter 3
  • Theodoridis and Koutroumbas, chapter 2, sections 2.5-2.5.4, 2.6

LINEAR CLASSIFIERS:

  • Schalkoff, chapter 4 (not discrete and binary classification problems)
  • Theodoridis and Koutroumbas, chapter 3, sections 3.1-3.5.2, 3.6 (SVM, not in 1st ed.)
  • A tutorial on Support Vector Machines for Pattern Recognition, http://www.kernel-machines.org/papers/Burges98.ps.gz

NONLINEAR CLASSIFIERS:

  • Theodoridis and Koutroumbas, chapter 4 sections 4.12-4.15, 4.17 (SVM, not in 1st ed.)

NEURAL NETWORK BASED METHODS:

  • Schalkoff, chapter 10, chapter 12
  • Theodoridis and Koutroumbas, chapter 4, sections 4.1-4.9, 4.11

UNSUPERVISED LEARNING AND CLUSTERING:

  • Schalkoff,
    • chapter 5;
    • chapter 13 (SOM)
  • Theodoridis and Koutroumbas,
    • chapter 11 (introduction), sections 11.1-11.1.1, 11.1.3;
    • chapter 12 (sequential algorithms), sections 12.1-12.4;
    • chapter 13 (hierarchical methods), sections 13.1-13.2.2, 13.2.5, 13.4;
    • chapter 14 (methods based on function optimization), sections 14.1, 14.5-14.6;
    • chapter 15 (methods based on competition), section 15.2.1, sections 15.3-15.3.2, 15.3.5-15.3.6

SYNTACTIC AND STRUCTURAL METHODS:

  • Schalkoff,
    • chapter 6 (formal languages);
    • chapter 7 (parsing) (NOT sections on Cocke-Younger-Kasami Parsing Algorithm, Augmented Transition Networks, Higher Dimensional Grammar, neither Stochastic Grammars);
    • chapter 8 (graphs) (to Extensions to the Elementary Graph Matching Approach, not including it);

All the subjects are not in one book; in the books the subjects have sometimes been presented with a much wider scope than in the course. It is sufficient to know the topics in the scope they were presented in in the lectures and exercises!



http://www.cis.hut.fi/Opinnot/T-61.231/2003/tenttivaatimus_en.shtml
markus.koskela@hut.fi
Thursday, 12-Aug-2004 13:41:54 EEST