Research And Evaluation On User Influence In Cobweb-Microblog Relationship Network

The most notable feature of the Web 2.0 era is user-centric, more focus on user interaction. Microblog is one of the representative technology in the Web 2.0 era, and it has simple publishing process and casual expression, so that users can freely participate in interactive in the online world. Because different personal values, different users will have different behavior, and will have different influence on other user in the same group. Microblog user influence refers ability that Microblog users affect other users in Microblog Relationship Network. Users are the core part of Microblog platform and researching on their influence in Microblog Relationship Network is conducive to the development of Microblog, providing a basis for further expansion, and providing new opportunities for different business applications.Microblog relationship network is one of social network, with the characteristic of small-world phenomenon a basic characteristic of complex network. In general, the connection between the Microblog users in the same small network is more closely, with sparse links to other small Microblog relationship network. We call a small Microblog relationship network which is normal user-based and with no obvious characteristics as Cobweb-Microblog Relationship network. Researching on the user Influence Evaluation in the Cobweb-Microblog Relationship network not only help to further research on Microblog user behavior, but also contribute to the research on Microblog Relationship network.In this paper, carefully sorting and analyzing of the existing researches on Microblog users influence, and summarizing the deficiencies they existing, we propose a comprehensive user quality-based user influence Evaluation model of the Cobweb-Microblog Relationship Network. Firstly, we analyzing the microblog user quality. We begin from the research on Microblog users’characteristics, analyzing the impact of active and passive behavioral characteristics of the users’on user influence, making three factors (user activity, coverage and Microblog Quality) to represent the user’s own quality based on seven selected user behavior features, and we put forward the concept and calculation method of the Users comprehensive quality. Then following the algorithm thinking of the famous link algorithm PageRank algorithm, we propose an user influence evaluation method in the Cobweb-Microblog Relationship Network combined with Users Comprehensive Quality.The data sets we’ll use in the experiments are real user data from Sina Microblog. We collect them using the user data acquisition supplied by cnpameng.com. Firstly, according to the model processes, we preprocess the raw data sets strictly, for example, removing the zombie user or the silence user, to ensure the quality of the experiment results. Then we visually analyze to the user data sets from different angles, to help users to understand the impact of User characteristics on user influence. Finally, we perform experiment with our method and analyze the results. We also make comparisons between the results and the results computed by other three methods, Followers assessment method, Centrality analysis method, and the traditional PageRank algorithm, and the results shows that our method is reasonable and effective.