摘要
为了提高元搜索引擎的查准率,提出一种改进的排序融合算法。首先,根据搜索返回结果中文档的位序以及包含该文档的成员系统数目计算文档初始评分;其次,引入BM25F算法模型计算文档的相似度;最后,增设域名缓存表统计文档的URL分值;综合上述三项计算值,得到文档的最终评分并作为排序依据。实验结果表明,所提出的优化算法显著提高了元搜索引擎系统的查准率。
To improve the precision of meta-search engine, an improved rank fusion algorithm is proposed. Firsly, the initial score of the document will be assigned according to the ranking position of the document in returned search result and the number of component engines that contains the document. Secondly, the BM25 F algorithm model will be introduced to calculate the similarity of the document. At last, an additional URL score of domain name cache table of the statistics document will be done. Integrating the above three calculated values, the final score of document will be derived and used as the basis of ranking. Experimental results show that the proposed optimisation algorithm significantly improves the precision of recta-search engine system.
出处
《计算机应用与软件》
CSCD
北大核心
2012年第10期188-190,226,共4页
Computer Applications and Software
基金
山西省回国留学人员科研资助项目(2011-028)
关键词
元搜索引擎
信息检索
排序融合
相似度
Meta-search engine Information retrieval Rank fusion Similarity
