| Summary: | 碩士 === 國立清華大學 === 電機工程學系 === 100 === In this thesis, we propose an algorithm that detects overlapping communities in networks (graphs) based on two simple node behavior models. The key idea in our algorithm is to find communities in a local agglomerative manner such that every community S has the following property: For each node i in S we have (1) the fraction of nodes in S that are
connected to node i is greater than a given threshold, or (2) the fraction of edges of node i that are connected to S is greater than another given threshold. Through extensive computer simulations of random graphs with built-in overlapping community structure, including the LFR benchmark random graphs and Erd¨os-R´enyi type random graphs, we show that our simple algorithm has excellent performance. Furthermore, we apply our algorithm to the
real-world network “Karate club” and show that the overlapping communities detected by our algorithm are very close to the known communities in this graph.
|
|---|
