计算机应用研究2018，Vol.35Issue(1)：121-125,5.DOI:10.3969/j.issn.1001-3695.2018.01.025
基于核心边的局部贪婪扩展与重叠性家庭群体识别
Core edge based local greedy expansion and overlapping family group recognition
刘荣辉 1高阳 2张湛梅1
作者信息
- 1. 中南大学商学院,长沙410083
- 2. 华为技术有限公司,广东深圳518129
- 折叠
摘要
Abstract
In order to support Telecom operators' precise marketing campaign regarding family user group,and solve the problem of detecting overlapping communities in large-scale weighted complex network,this paper put forward a core edge based local greedy expansion algorithm.Core edges were used as initial community seeds,two stages expansion were conducted based on weighted local fitness function.The algorithm prevented malformed expansion,defined edge weight likeness and weighted overlapping modularity,improved evaluation and processing of excessive similar communities,and enhanced overlapping communities division quality.Furthermore,based on telecom big data of China Mobile Co.,it constructed family member relationship model with logistic regression algorithm and detected family groups with CEBLGE algorithm.Through testing on sample dataset and the actual application in marketing campaigns,the model and algorithm are proved excellent.关键词
家庭群体识别/赋权复杂网络/重叠社团/局部贪婪扩展/核心边/逻辑回归Key words
family group detection/weighted complex network/overlapping communities/local greedy expansion/core edge/logistic regression分类
信息技术与安全科学引用本文复制引用
刘荣辉,高阳,张湛梅..基于核心边的局部贪婪扩展与重叠性家庭群体识别[J].计算机应用研究,2018,35(1):121-125,5.
