Many types of data have network-type structures, for example, social media networks, international trade networks, financial networks, terrorist networks. This is different from the classical numeric univariate/multivariate data since the network is characterized by nodes and edges. While modeling a network, what we try to characterize is: whether it is a dense network or a sparse one, how to measure the importance of nodes/players, how to compare two networks, etc. This would be the first part of the lecture.
Moreover, we are interested in the dynamics of a network, particularly in detecting changes in the network. This is a relatively new field, and professor Okhrin has a paper in this direction with coauthors from UCL and Uni Warwick. Thus in the second part of the talk, his work will be discussed.
University of Augsburg
Professor of Statistics
Experienced in: Data Analytics, Data Science, Business Forecasting, Econometrics, Time Series Analysis. Research: Machine Learning, Classification, Time Series, Econometrics; Statistical Process Control, Multivariate Data Analysis. Software used: R, C, Mathematica, LaTeX