Final written exam.
Basic concepts of causality according to Pearl and Spirtes are introduced, and the corresponding mathematical apparatus is developed. Learning algorithms based on the concept of graphical models and conditional independence are discussed. An important part is the practical estimation problem of learning causal graphs and statistical methods for estimating conditional independence. Since time series data are present in many applied sciences, special emphasis is placed on the challenges characteristic of time series. The course concludes with exemplary applications of methods to real data.
