Opinion community detection and opinion leader detection based on text information and network topology in cloud environment

With the rapid development of web technology, the social networks have become the largest information portals. In the social platforms, the text information can effectively reflect the user opinions or the public opinions for a certain entity, such as company, celebrity service, product and so on. Therefore, mining user opinions from social networks have become an imperative requirement for the service groups. In this paper, an opinion community detection method is proposed by considering the content similarity, the time similarity and the topology structure of users. The integrated similarity between two users, which includes the content similarity, the time similarity and the topology structure of users, is achieved. Then, based on the integrated similarities, the opinion communities are detected. Furthermore, in order to identify the opinion leader, an opinion leader detection method is proposed based on the user influence and emotional analysis. The users with the same topic form the opinion community. Meanwhile, a directed graph is created to formulate the interaction relationship between users in the opinion community. Then, the user influence model and emotional analysis model are presented. Moreover, the occurrence frequencies of the negative words are also considered in the emotional analysis model. Then, a model of influence value for each user in the opinion community is built. The user with the highest influence value is considered as the opinion leader. Finally, the performances of the proposed algorithms are evaluated in a distributed computing environment. Meanwhile, the extensive experiments are conducted. The results indicate that our proposed opinion community detection algorithm can effectively detect the opinion communities. Also, the proposed opinion leader detection algorithm can significantly identify the opinion leader in the social networks.