Abstract: In order to get better communities, the LSI method is used. Web page content and structure information are mined based on latent semantic, and the produced communities are integrated. Experimental results show that Web community integration algorithm of dual-view enhances semantic. The smaller communities is weakened. The results have higher accuracy than the ones based on a single link or content community mining algorithms. The search results have more remarkable improvement when user’s inputting keywords have not strong aspect-oriented purpose in information retrieval.

Key words: latent semantic, dual-view, Web community, integration algorithm

摘要： 为得到更好的Web社区划分，运用LSI方法，对Web页面的内容和结构信息分别进行基于潜在语义的社区挖掘，并对产生的社区进行集成。实验结果表明，Web双视图集成算法能够加强语义，使较小的社区划分被弱化，与单一的基于结构链接或内容的社区挖掘算法相比，具有更高的准确性。在信息检索的应用中发现，运用该算法检索特指性不强的关键词时，搜索效果有较明显改善。

关键词: 潜在语义, 双视图, 网络社区, 集成算法

CLC Number: 

• TP311