社群网络中基于链接分类的研究与应用

由于在互联网、电信、生物信息、社会网络分析等领域可获得的链接丰富的数据日益增多,链接挖掘已经成为数据挖掘的研究热点。基于链接分类是链接挖掘的一个重要方向。在此以电信领域用户通话特征数据为对象,研究了基于链接分类技术及其在电信客户流失预测的应用方法,在提取并分析大规模客户呼叫图的极大团、结点膨胀率、结点聚集度等与节点稳定性相关的链接属性及其时变特征的基础上,提出了一种适合海量数据的基于链接的电信客户流失预测算法。实验结果表明,此算法较传统分类算法能提高客户流失预测性能,实现了基于链接分类方法的成功应用。

Research and Application of Link Based Classification in Social Network

With the increase of structure data in the areas of internet,telecommunication,biology information and social network analysis (SNA),link mining has became an attracted research area in data mining.Link based classification is an important research direction in link mining.The research object is telecom customer calling character dataset,and the research is on the link based classification and its application in the telecom customer churn prediction.Based on the calculation and analysis of the structure of the large scale call graph,such as clique,expansibility,centrality,and their correlation with the stable state of the vertex.An algorithm fits for the telecom customer churn prediction in large scale dataset is proposed.The result of examination indicates that this algorithm promote the capability of the customer churn prediction model,it's surely a successful application of link based classification.