PhD Seminar: "Simultaneous Anomalous Network Clustering"

Speaker: Soroosh Shalileh, second-year PhD student, School of Data Analysis and Artificial Intelligence, Faculty of Computer Science
Where: Faculty of Computer Science, Kochnovskii proezd 3, room 205 
When: May 13, 18.10–19:30 

Graphs are a powerful mathematical tool to model complex systems. They can be utilised in various applications from microbiological systems to social interactions. One approach to analyse graphs is by applying cluster analysis methods. To this end, the variety of approaches have been proposed. Most of them are based on analysing only the graph structure, but some recent approaches also take into account node attributes. In this talk, we propose a Simultaneous Anomalous Network Clustering (SANC) method based on graph structure and node attributes. One of the advantages of the proposed method is the ability to adjust the impacts of the graph structure and node attributes in clustering process. We evaluated the performance of the proposed method on synthetic data and real-life data sets.