基于超链接和内容相关度的检索算法  

Retrieval algorithm based on hyperlinks and content similarity

作　　者：张娜[1] 张化祥[1] 

机构地区：[1]山东师范大学信息科学与工程学院,山东济南250014

出　　处：《计算机应用》2006年第5期1171-1173,共3页journal of Computer Applications

摘　　要：在网络环境下,经典的链接分析方法(HITS算法)过多的关注网页的权威性,忽视了其主题相关度,易产生主题漂移现象。在简要介绍HITS算法的基础上,分析了其产生主题漂移的原因,并结合内容相关度评价方法,提出了一种新的搜索算法———WHITS算法。实验表明,该算法挖掘了超链接间的潜在语义关系,能有效的引导主题挖掘。Under the circumstances of web, classical hyperlink analysis algorithms（such as HITS algorithm） mainly focused on the authority of a web page rather than its topic, so it was easy to drift away from the mining topic when traversing the hypedinks. The cause of topic drifting away in HITS algorithm was analyzed. By combining the topic analysis method with the content relevance evaluation, a novel web information retrieval algorithm - WHIrlS was presented. Experiment results show that WHIrlS focuses on mining the potentially semantic relationship between hyperlinks and performs quite well in the toplc-specific crawling.

关 键 词：HITS算法 超链接 向量空间模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]