Machine Learning

The course is concerned with deeper explanation of traditional machine learning models and algorithms. Particular interest, besides the main principles of these algorithms and their theoretical background, will be devoted to hyper-parameters of various algorithms such as their meaning and tuning. The pros and cons of these algorithms w.r.t. various application domains and prediction tasks will be discussed, too. Main topics of the course include decision trees, support vector machines and kernel methods, graphical and probabilistic models, neural networks, factorization techniques, semi-supervised learning, ensemble techniques, bagging, boosting, time-series and text-mining.

Major topics:

• decision trees;
• support vector machines and kernel methods;
• graphical and probabilistic models;
• neural networks;
• factorization techniques;
• semi-supervised learning;
• ensemble techniques, bagging, boosting;
• time-series mining;