互联网天地
互聯網天地
호련망천지
HULIANWANG TIANDI
2015年
8期
33-41
,共9页
微博网络%社区发现%主题模型%JS距离%CNM算法
微博網絡%社區髮現%主題模型%JS距離%CNM算法
미박망락%사구발현%주제모형%JS거리%CNM산법
microblogging network%community discovery%topic model%Jensen-Shannon divergence%CNM algorithm
提出一种将用户关系和用户兴趣两者融合的社区发现算法——JSCNM算法，然后利用JS距离公式计算用户兴趣相似度距离，使用相似度代替模块度对用户关系和用户兴趣双内聚处理，发现用户社区。最后通过爬取新浪微博4个数据集进行实验操作，并对照CNM算法和GN算法，结合社区评价标准模块度的大小，证明该算法能够更准确、合理地发现用户社区。
提齣一種將用戶關繫和用戶興趣兩者融閤的社區髮現算法——JSCNM算法，然後利用JS距離公式計算用戶興趣相似度距離，使用相似度代替模塊度對用戶關繫和用戶興趣雙內聚處理，髮現用戶社區。最後通過爬取新浪微博4箇數據集進行實驗操作，併對照CNM算法和GN算法，結閤社區評價標準模塊度的大小，證明該算法能夠更準確、閤理地髮現用戶社區。
제출일충장용호관계화용호흥취량자융합적사구발현산법——JSCNM산법，연후이용JS거리공식계산용호흥취상사도거리，사용상사도대체모괴도대용호관계화용호흥취쌍내취처리，발현용호사구。최후통과파취신랑미박4개수거집진행실험조작，병대조CNM산법화GN산법，결합사구평개표준모괴도적대소，증명해산법능구경준학、합리지발현용호사구。
A kind of community discovery algorithm which was named JSCNM fused the relationship between user links and user interests was proposed. JS distance formula was used to calculate the user interest similarity distance, and the user community was discovered by using the similarity degree of relationship between user links and interests instead of modularity. Finally, through CNM and GN algorithm by using four experimental data set for operation and combined with modularity the community evaluation standard, JSCNM algorithm was proved to be better in finding more accurate and reasonable user communities.
