Multi - Relationship Community Detection Based On Shaply Value

People’s interaction and cooperation creat the social intercourse.Variety of social activities make the life of people more colourful. Social analysis is the study of interaction between people and cooperation patterns of behaviour, has always been a hot area of research, communities detecting is an important direction of social analysis. With continuous upgrades of communications technologies, communication between people become more convenient. Emergence of Web2.0has brought revolutionary changes to people’s social activities, along with the popularity of representative social media such as Twitter, Facebook, renrennet, blurring the gap between communication and virtual communication. Analysis results of online social networks, directly or indirectly reflect the true social, reflected the growing importance of Web Analytics, Web Analytics on combating crime, social studies, the spread of the virus, e-commerce, has played an important role.Current community detecting research focused on a single social network, gives results only reflecting a kind of relationship of social network. In the real world, the relations between man and man are complicated. In General, the number of relation is greater than one.Each relationship represents a network, the network performance varies. These relationships interact with each other, which are true reflection of people’s social actions. Multi-relationship social network provide richer interaction information, more genuinely social scene. Therefore, multi-relationship social network research make the sence.Our work propose community detecting algorithm based on shapely value (SHMRCD), the algorithm make efforts on fair interest allocation for each individual,which model the community formation process as each individual rational choice process. Combining multy social networks according user’s query without prior setting the number of communities,our algorithm automaticly generate community dividing results. Being different from the single relationship netwok, that individual rational choice in multiple relations made the decision reflecting the combined effect of many relationships.We extend the module method to evaluate the quality of the community partition result. Lots of experiments show that,to a certain extent,the communitie detection algorithm for multi-relationship social networks is applicable.