Syllabus
Linear regression and instance based learning as "extremal points" in the space of models,
Undirected graphical models, Gaussian processes and Bayesian optimization, basis expansion and regularization (smoothing splines and other methods), logistic regression, generalized additive models, model assessment (crossvalidation, one-leave-out, analytical criteria) decision trees, prunning, missing values, rule search PRIM, model averaging, boosting, random forest, support vector machines,
Bayesian learning, EM algorithm introduced on an clustering example, unsupervised learning - market basket analysis, clustering k-means, k-medoids, hierarchical clustering,
Inductive logic programming.
Annotation
The aim of the course is to introduce machine learning as important and in this time very vital field developing in the close connection with artificial intelligence. The course gives a survey of basic branches of machine learning (supervised inductive learning, reinforcement learning, unsupervised learning and knowledge in learning), main problems and methods and some typical algorithms.