Abstract: This paper proposes a novel query expansion method using Local Context Analysis(LCA) based concept tree pruning. It extracts the suggested terms from initial documents which are retrieved for the original query by LCA method, and uses these terms to prune the concept tree built by lexical database, to add new terms, and to recalculate the weight of expanding terms. Experiments show that the combined expansion yields significant improvements over existing algorithms under the same experimental condition.

Key words: query expansion, concept tree, Local Context Analysis(LCA)

摘要： 介绍一种局部上下文分析(LCA)剪枝概念树的方法。利用LCA方法初次检索的与原查询最相关的文章作为备选扩展词的来源，用扩展词剪枝语义词典构造的概念树，补充概念树上不存在的新词，并重新计算扩展词权重。实验表明，在相同的实验条件下该扩展方法查询性能有较大的提高。

关键词: 查询扩展, 概念树, 局部上下文分析

CLC Number: 

• TP311