摘要
分析了Web页面主题的分布的特点,对经典的页面排序算法进行了探讨,提出了一种基于内容和超链接分析并结合用户点击行为的相关性排序算法。该算法考虑了超文本标记、锚文本、文本内容等对相关性的影响,引入动态比较矩阵来计算相应的权重系数,能够客观分析网页所包含的主题信息,使检索结果排序更合理。实验表明,该算法能有效提高查准率,较好地解决了主题的漂移现象,且具有较好的性能。
The distribution characteristic of Web page topic is analyzed,the classics ranking algorithm of Web pages is discussed,one kind of correlation ranking algorithm of Web page topic is proposed,which is based on the content analysis and hyperlinks analysis,associated with the user clicks.Considering the hypertext marks,anchor text,text content etc. on the correlation, and introducing the dynamic comparison matrix to calculate the corresponding weighting coefficients,the algorithm can analyzes the topic information in the Web page objectively,and guarantee retrieval result rationality.Experiments indicate that,it can enhance the retrieves accuracy ratio effectively,and solve the topic drift well.This algorithm has the good performance.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第23期149-151,157,共4页
Computer Engineering and Applications
基金
西南大学发展基金资助项目
