The Discovery of Micro-Blog Users' Interest Groups Based on SSLOK-Means Clustering Focus on Big Data Improved by Calinski-Harabasz

As a data mining algorithm, the clustering has been widely used in the discovery of micro-blog users' interest groups and provides a new idea to mine users' information for the micro-blog operators. The SSLOK-means algorithm newly proposed can solve the drawback of K-means algorithm when clustering the big data. However, the number of clusters still needs to be set manually in advance, which brings difficulties for using. In this paper, an algorithm of semi supervised labels one scan K-means improved by Calinski-Harabasz function is presented and applied to subdivide users and to discover the interest groups. The micro-blog users' data of Sina have been grabbed to do clustering analysis by the new algorithm. Finally, six different types of interest groups including knowledgeable type, material type, IT type, business type, interconnected type and balanced type have been found. The targeted individual recommendations were given for each type of interest group. The research can provide a support for the optimization of personalized service about micro-blog platform and the earnings promotion of micro-blog marketing, which has important theoretical significance and practical value.