面向DBWorld数据挖掘的学术社区发现算法
发布时间：2018-11-12 13:21
【摘要】：针对传统社区发现算法多数是基于单一关系的同构学术社会网络,而包含多种关系的异构学术网络社区发现算法还不多的情况,提出一种基于FCM(fuzzy C-means)和结构洞的学术社区发现算法——HAFCD算法。从构建基于DBWorld邮件数据的异构学术网络出发,通过分析异构网络中的多种关联关系和节点内容的相似性,提出改进的语义路径模型,计算评审人间的相似度。基于此,该算法根据结构洞越少、网络闭合性越高这一事实,将结构洞理论融入FCM算法进行异构学术社区发现。通过与现有的谱聚类和路径选择聚类算法进行实验比较表明,该算法具有较好的计算效果。
[Abstract]:Most of the traditional community discovery algorithms are isomorphic academic social networks based on a single relationship, but there are not many heterogeneous academic network community discovery algorithms including multiple relationships. This paper presents an algorithm for discovering academic community based on FCM (fuzzy C-means) and structure hole, which is called HAFCD algorithm. Based on the construction of heterogeneous academic network based on DBWorld email data, this paper proposes an improved semantic path model to calculate the similarity between reviewers by analyzing the various association relationships and the similarity of node content in heterogeneous networks. Based on the fact that the fewer the structure holes and the higher the network closeness, the structure hole theory is incorporated into the FCM algorithm for the discovery of the heterogeneous academic community. Compared with the existing spectral clustering and path selection clustering algorithms, the experimental results show that the proposed algorithm is effective.
【作者单位】： 上海理工大学光电信息与计算机工程学院;
【基金】：上海智能家居大规模物联共性技术工程中心资助项目(GCZX14014) 沪江基金研究基地专项项目(C14001) 国家自然科学基金资助项目(61003031)
【分类号】：TP311.13
[Abstract]:Most of the traditional community discovery algorithms are isomorphic academic social networks based on a single relationship, but there are not many heterogeneous academic network community discovery algorithms including multiple relationships. This paper presents an algorithm for discovering academic community based on FCM (fuzzy C-means) and structure hole, which is called HAFCD algorithm. Based on the construction of heterogeneous academic network based on DBWorld email data, this paper proposes an improved semantic path model to calculate the similarity between reviewers by analyzing the various association relationships and the similarity of node content in heterogeneous networks. Based on the fact that the fewer the structure holes and the higher the network closeness, the structure hole theory is incorporated into the FCM algorithm for the discovery of the heterogeneous academic community. Compared with the existing spectral clustering and path selection clustering algorithms, the experimental results show that the proposed algorithm is effective.
【作者单位】： 上海理工大学光电信息与计算机工程学院;
【基金】：上海智能家居大规模物联共性技术工程中心资助项目(GCZX14014) 沪江基金研究基地专项项目(C14001) 国家自然科学基金资助项目(61003031)
【分类号】：TP311.13
【相似文献】
相关期刊论文 前10条
1 陈庆章;组建异构网应着重考虑的问题和产品现状[J];计算机时代;1994年03期
2 汪芸,顾冠群,谢俊清,兑继英,孙昌平;异构网络集成方法研究[J];计算机研究与发展;1997年03期
3 马义忠;杨红旗;高彦;専秋峰;;基于移动Agent异构网络管理的分析与设计[J];微计算机信息;2008年33期
4 黄川;郑宝玉;;多无线电协作技术与异构网络融合[J];中兴通讯技术;2008年03期
5 吴蒙;季丽娜;王X;;无线异构网络的关键安全技术[J];中兴通讯技术;2008年03期
6 李R，
本文编号：2327240
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2327240.html
