电信大规模社交关系网络图数据挖掘研究  

Research on Large-Scale Social Network Graph Data Mining in Telecommunication

作　　者：刘丽娇[1] 陶俊才[1] 肖晓军 卢宇 

机构地区：[1]南昌大学信息工程学院计算中心,南昌330029 [2]广州优亿信息科技有限公司,广州510630

出　　处：《电信科学》2015年第1期23-31,共9页Telecommunications Science

摘　　要：电信技术的发展使得通话网络几乎覆盖了整个社会,对电信运营商来说,这个庞大的社会网络蕴藏着巨大的商机。尤其在日益激烈的市场竞争中,客户关系的维护、客户潜在价值的挖掘以及有针对性地对客户进行营销等对电信图数据的挖掘,受到了强烈关注。通过对比,研究了当前针对类似电信网络的大规模社交关系网络图数据进行挖掘的分布式和单机图计算工具和框架,并对单机图计算工具Graphchi的单机处理数据能力进行了可行性和可用性测试,同时采用单机Graphchi对电信社交关系网络的数据进行了图挖掘研究和应用。With the development of the telecommunications technology, telephone call network almost covered the whole society; the huge social network presents tremendous opportunities for telecommunication companies. Especially in the increasing fierce market competition, more and more telecommunication companies are concerned on maintaining customer relationship, mining customer potential value and marketing suitable target customers. The distribute and stand-alone graph computing frameworks and tools were researched for mining large-scale social network graph data like telecommunication network, and at the end the feasibility and usability of Graphchi in large-scale graph datasets were verified and a graph mining application was made with telecommunication data using stand-alone graph data processing tool--Graphchi.

关 键 词：电信图数据 大规模图数据挖掘 Graphchi 

分 类 号：TN91[电子电信—通信与信息系统]