1. komponenta

Lecture type	Total
Lectures	30
Exercises	15

* Load is given in academic hour (1 academic hour = 45 minutes)

COURSE AIMS AND OBJECTIVES: Course aim is to present the key algorithms and theory from the core of machine learning and to learn how to construct the computer programs that automatically improve with experience (data).

COURSE DESCRIPTION AND SYLLABUS:
Introduction to machine learning. Solving classification, clustering and regression problem based on machine learning methods. Concept learning. Generalization. Supervised unsupervised and reinforcement learning.
Percepron. Perceptron learning. Linear regression and least squares methods. Gradient descent and Delta rule. Kernel perceptron.
Neural networks. Neural network learning. Multilayer networks and backpropagation algorithm. Neural networks for regression and clussification problems.
Classification. Fisher's linear discriminat analysis. Logistic regression. Support vector machines. Examples of use in bioinformatics and automatic document classification.
Non-parametric learning techniques. k-nearest neighbor algorithm. Decision tree learning.
Bayesian learning. MAP i ML hypotheses. MDL principle. Bayes optimal classifier and Bayes naive classifier. EM-algorithm.
Unsupervised learning and data mining. Hierarchical Clustering. k-means clustering. Dimensionality reduction and principal component analysis. Applications and examples.
Statistical learning theory. PAC learnability. Vapnik - Chervonenkis dimension.

1. T. Hastie, R. Tibshirani, J. H. Friedman: The Elements of Statistical learning
2. T. M. Mitchell: Machine Learning
3. V. Vapnik: The Nature of Statistical Learning Theory
4. N. Cristianini, J. Shawe - Taylor: An Introduction to Support Vector Machines

Izborni računarski predmet 1, 2 - Regular study - Mathematics and Computer Science Education

Izborni računarski predmet 1, 2 - Regular study - Mathematics and Computer Science Education