基于概念空间方法的信息检索技术研究

A Study on Information Retrieval Based on Concept Space Approach

为了解决词汇差异问题,词表构造在信息检索系统中有着重要意义。概念空间方法是利用计算机自动构造概念语义网络(词表)并以此为基础进行概念检索的一种方法。由词语作为语义网络的节点,词语之间的关联权重以一个给定文档集合中词语的共现率来计算,其大小代表它们之间的相似性。检索时系统采用人工智能方法激活与检索入口词相关的术语或概念,为用户提供交互式的检索用语建议。方法的具体步骤包括文档和对象列表收集、对象过滤和自动标引、共现分析和联想检索四个阶段。这种方法多用于英文检索系统,但对我国的信息检索系统也有重要的借鉴意义。

The construction of thesaurus is significant in the information retrieval system for the sake of solving the vocabulary difference problem. This paper introduces Concept Space approach, which is an approach to constructing the semantic network, i.e. thesaurus by computer and applying it to information retrieval. The specific steps of Concept Space include document and object list collection, object filtering and automatic indexing, co-occurrence analysis, and association retrieval. This approach has been employed by many English information retrieval systems, and could presumably be replanted to Chinese information retrieval systems.