Detekce lokálních komunit v komplexních sítích

Abstract

This thesis deals with problems of local overlapping community detection in complex networks. During the last ten years complex network analysis has become very popular discipline thanks to the growth of social networks and data availability on the WWW. Big effort was involved especially in the field of community detection in complex networks. There are few different approaches to the community detection, but they can be divided into two main branches. Either we can detect global or local communities. These are further divided depending on fact if they allow vertices to belong to multiple communities or not. We will describe the differences and specifics of each method, but we will focus mainly on local methods. Two methods which variations were implemented will be discussed in detail. These are Clique percolation method(CPM) and Connected iterative scan (CIS). In the end of this thesis we will discuss performed experiments comparing implemented methods and their results.