Based In The Six Degrees Of Separation And Centrality To Identify The Key Figure In The Micro-blog

With the evolution of the Internet and technology, we have ushered in the era of Web2.0,as a new communication carriers, Micro-blog rise rapid and are subject to people’s attention.Compared with the traditional media, Micro-blog has many features，such as large amount ofinformation, fast information spreading, strong instantaneity, which make it become a newplatform for people to exchange and transmit information, but also to the spread of rumors ofnew channels. Faced with the amount of data Micro-blog which generated billions of dollarsevery day. How do we deal with these data and recognize the key nodes in Micro-blog timelyand effectively are challenges that we are going to face. This paper analyzes the basicfunctions, information characteristics and propagation characteristics of the Micro-blog, andon the basis of the existing centrality and clustering classification algorithm which at homeand abroad, we focus on the technology of identifying network key node of Micro-blog, withthe balance processing in time and the accuracy, the proposed solutions as follow:（1） Analyzed the characteristics of the Micro-blog, and combined with the algorithm forCentrality Dependences, we proposed the Interactive Index and the Affective Index, thencombined the two, finally found the key index that apply to the microblog to recognize thekey figure in the Microblogging Network.（2） Faced with the vast amounts of data which are generated every day of Micro-blog,we combined the idea of six degrees of separation theory, use the SPLINE algorithm, wefinished the set of nodes again which Sub-group size is too large, remove the part of anytwo nodes distance greater than5. After this step, we get a more closely linked set of thenodes.（3） Against the closely linked set of nodes, we analyzed the hierarchical clusteringalgorithm, use the CNM algorithm to cluster the data set, and gathered the high similaritymicroblog, formed a closely linked dataset.（4） According the important node set of the proceeds of the third step, use the algorithmof the key node for it, identify the important node in the concentration of key nodes effective.Finally, verify the proposed method, we use the method to take an algorithm program,Micro-blog network identification of the key nodes on the NLPIR microblogging contentCorpus. For the results of the experimental, we analyze it, prove that our approach is efficientand practical.