元搜索引擎的结果去重及排序研究

Meta Search Engine Results to Weight and Sorting of

结果去重和排序是提高元搜索引擎结果质量的两个关键问题,文章分析余弦相似度、基于TF-IDF的文本相似度三种去重算法,利用URL、标题和计算摘要相似度三方面去重;研究了Board排序、星星排序、轮询法、位置排序和概念可行度对检索结果的影响,提出了一种综合排序算法。实验结果表明,综合排序算法在准确率、召回率等方面都优于其他算法。

Go heavy and sorting are two key issues to improve the quality of the results of the meta-search engine,the article analyzes the cosine similarity,three kinds of text similarity based on TF-IDF weight algorithm,using the URL,title,and calculation of summary similarity toweight;Board sort stars sort,the polling method,location,sort,and the concept of feasible search results,a comprehensive sorting algorithm.The experimental results show that the integrated sorting algorithm accuracy,recall rate of better than other algorithms.