Home
 
 
Course Description
 
ΥΠ2 - Pattern recognition
Code: ΥΠ2
Title Course: Pattern recognition
Semester: 1st
Category: REQUIRED
Lecture hours: 3
Instructor: Perantonis (Demokritos)
Course web page: http://eclass.uoa.gr/courses/D449/
Course Description:
Classification based on Bayes decision theory (basic principles; Bayes classifiers for normal distributions; estimation of probability density: maximum likelihood estimation, maximum a posteriori probability, maximum entropy). Bayesian networks. Linear classifiers (single layer perceptron, LMS algorithm, support vector machines). Non-linear classifiers (decision trees, multilayer perceptrons, radial basis functions, non-linear support vector machines). Context based classification (Markovian chains, Viterbi algorithm, hidden Markov models). Introduction to feature selection and extraction (statistical hypothesis testing, search methods, principal component analysis, linear discriminant analysis, moments, discrete Fourier transform, wavelets). Introduction to clustering (examples of clustering algorithms: serial algorithms, isodata, self organizing maps). Pattern matching (Bellman's optimality principle, dynamical programming, Levenshtein distance). [Previous page]
 
Sitemap | Contact | Credits | Rss
National and Kapodistrian University of Athens Department of Informatics and Telecommunications