Similarity-based Complex Network Community Detection Research

With the development of information technology, human society has stepped into network era.Complex network system can be seen everywhere. Community structure is an important characterand hot research direction of complex network. Division of network community contributes to abetter understanding of community structure and predicts the behavior of complex network, andfurthermore, it has great application value in social network, information recommendation andprecision marketing, and so on. Aim at community division algorithm of complex network, work ofthis paper can be summarized as follows:1. Build a LR-S hierarchical network based on real-time communication network and personalnetwork. There are two evaluation factors for measuring relationship of nodes in the network:activeness of communication network and the shortest path length of personal network.2. Propose a SA-LEN evaluation function of similarity degree of nodes based on the twocharacters.This function combines physical distance between humans in real-world network with lo-gical interpersonal relationship to make the description of node s similarity degree better.3. Propose a SA-LEN community detecting algorithm based on SA-LEN similarity degree ofnode. This algorithm is tested on nature networks and computer generated networks, and wasproved to be effectiveness.4. Present a software implement of complex network simulation which is designed by myself.The detailed design methods and the running effects are given in the thesis.