Research And Implementation On The Graph Mining Technology Of Weibo Community

In recent years, Weibo as a popular SNS, with its large amount of information, high openness and low entry threshold, making the number and activity of users maintained rapid growth. With the increase of the number of users, the Weibo’s relationship is becoming more and more complicated. Through blogs, topic or focus, and etc, the users of Weibo formed a community with some characteristic, which is include similar hobbies, friends or value, etc. When dealing with a large number of relationship, the data mining methods of relational database have been difficult to meet the demand of practical application.In this paper, using the graph data mining techniques and tools to Weibo user correlation processing, can find the characteristics of user groups, and can effectively user recommendation, event analysis, product promotion and marketing, etc. For the problem of the Graph Mining in Weibo community, this paper mainly do the several aspects works:Firstly, Based on graph data mining technology and build a weibo user relationship structure model. According to the characteristics of weibo relationship structure, we propose two kinds of random walk method to generate pattern, and mining the frequent pattern in relational structure model by the pattern support, then clustering the pattern by the structural similarity between two pattern, and find the characteristics structure finally. Finally through the use of graph database to deal with sina weibo focused on relational data, verification and comparison of the two methods of processing efficiency and effectiveness.Secondly, Weibo is widely use the labels to mark users, this paper will use similar label algorithm of graph mining into frequent pattern mining, to improve the effectiveness of finding the characteristics structure, and through the experimental analysis, finally shows the evaluation results of the similar label graph mining and Weibo characteristics structure.Finally, according to the requirements analysis, we proposed and implemented a scalable architecture of Weibo community detection, and introduces the service strategy.