基于潜在语义索引的超链接分析模型

ranking model for the Web is proposed to bring more accurate and rational order for Web documents on the World Wide Web. The model combines hyperlink analysis based on Markov chain and content or topic analysis based on Latent Semantic Indexing and hierarchical clustering. This model is sensitive to newly published Web and can be used to combat content spamming and in-link spamming in web search engine, to extend user query through user relevance feedback and also to topic of user query sensitively search in context.