A Study on Meta-search Based on Abstract Analysis
基于网页摘要分析的元搜索引擎研究*

Facing too much results returned by different search engines, this paper designs a meta- search in order to fully utilize the results. It introduces the basic structure of meta - search and methods of results merging. And employs the techniques of statistics to study the relevance between title and page content, relevance between abstract and page content, and then fixes the weights of title and abstract when judge the relevance of search results. The experiment proves the improvement of average veracity comparing with the member search engines.