Analysis of Social Group Dynamics

TL;DR

This paper analyzes the GED method for social group evolution detection, comparing it with other techniques in terms of accuracy, efficiency, and flexibility, across various social group scenarios.

Contribution

It provides a comprehensive evaluation of the GED method against other approaches, highlighting its strengths and limitations in social group analysis.

Findings

GED shows competitive accuracy in group evolution detection.

GED's computational cost varies with different user importance measures.

The method is flexible and adaptable to different social group types.

Abstract

In this thesis the method for social group evolution discovery, called GED, is analyzed. Especially, GED method is compared with other methods tracking changes in groups over time with focus on accuracy, computational cost, ease of implementation and flexibility of the methods. The methods are evaluated on overlapping and disjoint social groups. Finally, GED method is run with different user importance measures.