The Research On Local Community Detection Based On Relative Intimacy

The development of information processing technology keeps promoting the analysis of complex networks. Many important characteristics of complex networks was discovered, such as "small world" and "scale free", which helps the understanding of network structure and network evolution process. Community is important in networks, and is valuable in some research areas such as information technology, social science, and life science. Community detection attracts more attention because the networks are large, complex, and difficult to be verified.This paper focus on the issues of local community detection, and the details are as follows:1) The metric of "Relative Intimacy" is proposed. This metric defines the differences of arbitrary two nodes, which indicates the difference of node social status. All the proposed algorithms are based on the metric.2) A local community detection algorithm based on "Relative Intimacy" is proposed. A new fitness function is proposed based on a classic expanding model. The Experiments show that the algorithm is fast and efficient.3) A new local community detection algorithm is proposed which focus on Massive Call Graphs. MCG Network is a special Complex Network, which reflects real human activities. Local community detection on MCG Network is meaningful in many fields, such as business marketing, crime prevention etc.