Social Relation Network Feature Analysis Based On Weibo

With the rapid development of web2.0, web2.0 as Sina Weibo tangible information carrier also showed rapid growth. Different regions, different personalities user groups needs to show up is a big difference. How to deepen the differences and needs of the user community in the understanding of user behavior, the identity of reasoning, analysis and precision marketing groups. To the development of enterprises provides real user demand information to the government to understand the source of livelihood to provide information platform, while enterprises to microblogging users to provide personalized services to help solve modern people choose phobia.This study based on April 18, 2014 sina weibo user data, with the aid of sina weibo API(Application programming interface) interface, combined with the use of web crawler technology grasping its weibo user information, mainly USES the R statistical language software to realize BP neural network classification algorithm analysis weibo user attribute Tag content, categorizing users Tag, reveal the hot topic of users in the label.At the same time, the paper mainly compares the several clustering algorithms: K-means, the hierarchical partitioning clustering DIANA, agglomerate hierarchical clustering AGNES and CLARA algorithm for the realization of the user groups cluster analysis.Finally realizes the research object is the basic feature of statistical analysis, the numerical character of user, user tag preference analysis and clustering analysis of user groups.Sina weibo, the most representative of social network, it has broad prospects in the field of application and research, especially in the recommended information, e-commerce, journalism and communication and human dynamics, etc.Research is based on the purpose of sina weibo scattered integrating original data information, and found that the potential of infinite value in application and research, so as to bring value to the development of society.Finally with the help of classification results can show the weibo user attention to provide personalized service, and the results of the study to the enterprise to provide strategic decision information support, help enterprises to find more field, the user's interests accurate customer demand, and promote enterprise to interact with customers in real time.