分布式检索中查询结果合并方法研究  

Research on Search Results Merging Methods in Distributed Information Retrieval

作　　者：刘婧[1] 袁可[1] 蒋伟[1] 

机构地区：[1]西南大学计算机与信息科学学院,重庆400715

出　　处：《情报学报》2009年第5期685-688,共4页Journal of the China Society for Scientific and Technical Information

摘　　要：查询结果合并是分布式信息检索中的一个重要步骤，其合并方法的选择直接影响检索结果的质量。本文首先对两种查询结果合并算法，即经典的CORI算法与新近提出的回归分析与选择下载相结合的Hybrid算法的基本原理进行了讨论研究，然后通过实验对这两种算法的性能进行了深入的比较分析。在实验过程中，采用平均准确率指标对检索结果进行评价。通过比较两种合并算法产生的平均准确率来评价二者的性能。结果表明，在不同的实验环境下，新的Hybrid算法的性能都要优于CORI算法。选择Hybrid算法进行查询结果合并能够取得令人满意的结果，比较适合作为分布式检索的查询结果合并算法。Search results merging is an important process of distributed information retrieval and the selection of merging methods directly influences the quality of retrieval results. This paper first discusses the main principle of two results merging methods, the well-know method CORI and the new proposed Hybrid algorithm that combines regression analysis and selective download, then compares the performance of these two algorithms by experiments. In the experiments, precision is used to evaluate search results. By comparing the precision that two merging algorithms generate, we can evaluate the performance of them. Finally, results demonstrate that in different test environments the new Hybrid algorithm gains better performance than CORI. Selecting Hybrid algorithm to merge search results can achieve a promising result. Therefore, Hybrid is more appropriate to be the search results merging algorithm in distributed information retrieval.

关 键 词：分布式信息检索 结果合并 HYBRID CORI 

分 类 号：G354[文化科学—情报学] F231.5[经济管理—会计学]