Abstract: Most retrieval systems ignore the anchor text, which is highly relevant to the page content in Web information retrieving. This paper proposes a method to improve the retrieval accuracy. It makes two indices, one for page content and the other for anchor text. A parallel retrieval strategy is utilized for the two indices. Experimental results show that this method is efficient for the special structure document collection.

Key words: anchor text, parallel retrieval, information retrieval

摘要： 进行Web信息检索时，页面中的锚文本与正文存在较大相关性，多数检索系统忽视了锚文本对页面正文的贡献。该文提出一种提高检索精度的方法，为文档集建立一个基于页面正文的索引和一个基于锚文本的索引，对其采取并行检索策略。实验结果表明，该方法可以有效处理特定结构的网页集。

关键词: 锚文本, 并行检索, 信息检索

CLC Number: 

• TP391