基于内容分析的短信种子客户挖掘模型与算法  

SMS seed customers mining model and algorithm based on content analysis

作　　者：黄志超[1] 陶俊才[1] 高胜保[2] 

机构地区：[1]南昌大学信息工程学院计算中心,江西南昌330029 [2]中国电信股份有限公司江西分公司,江西南昌330029

出　　处：《电信科学》2016年第2期106-111,共6页Telecommunications Science

基　　金：国家自然科学基金资助项目(No.61262049)~~

摘　　要：为从海量的短信记录中挖掘短信种子客户,控制种子短信的传播路径,提高其传播效率,提出了一种基于内容分析的短信种子客户挖掘模型与算法。首先通过分析客户转发短信的兴趣性、随机性、单向性特征,构建客户转发短信的树型模型;其次,通过定义和应用综合评价函数生成优化的种子客户挖掘模型,并基于亲密群概念实现短信种子客户的挖掘;最后,使用电信运营商的实际数据进行实证分析,验证了上述模型与算法的有效性。In order to mining SMS （short message service） seed customers from massive text messages, control the spread of the seed messages path and improve the efficiency of its spread, a SMS seed customers mining model and algorithm was proposed, which was based on content analysis. First of all, by analyzing the interest, randomness and one-way characteristics of customer forwarding messages, the tree model Of customer forwarding messages were constructed. Secondly, the optimal seed customers mining model was generated by definition and application of comprehensive evaluation function, and SMS seed customers mining was realized based on the concept of close group. Finally, by analyzing the actual data from telecom operators, the effectivity of the model and algorithm was verified.

关 键 词：短信种子客户 内容分析 挖掘模型 挖掘算法 亲密群 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]