基于概念的信息检索模型研究

Research on the Concept-based Information Retrieval Model

随着Internet的迅速发展 ,WWW已经成为世界上最大的信息库 ,它正日益改变着人类的生活方式 .然而 ,由于WWW信息资源庞大 ,结构复杂 ,如何高效地从中找到需要的信息 ,已经成为困扰网络用户的一大难题 .许多著名的站点 ,如Yahoo ,AltaVista ,Infoseek均使用基于关键字的搜索引擎 ,存在明显的缺陷 ,当查询用的关键字与目标文档尽管语义相同 ,但用词不一致时 ,将检索失败 ,导致召回率很低 .提出一个基于概念的信息检索模型 ,它不是以关键字为核心 ,而是以概念为核心来实现信息检索 .着重介绍了基于概念的信息检索模型的设施。

With the rapid development of Internet, World Wide Web has become a large information resource of the world. It changes the life mode of human being. However, because the resource is very big, and the structure is very complex, how to search and retrieve information efficiently and effectively becomes an important problem. The traditional search engines, such as Yahoo, AltaVista, InfoSeek are keyword-based search engine. They have an obvious default in common. When the word or phrase in the query is different from those used in the material you needs, searching with failed though these have a common sense. This leads to low recall. In this paper, we'll present a concept-based searching engine model. It uses concept instead of keyword as the kernel to complete the information search. This paper briefly introduces the facilities, methods and tools of the Concept-based Information Retrieval Model. The main contents of this paper are (1) to design and build the concept lexicon that supports the mapping between term and concept. At last these concepts can be found in the concept-tree;(2)to design and build the concept-tree that expresses the hierarchy of the knowledge and the relation among concepts. The concept lexicon and the concept-tree constitute the meta-knowledge of the model. The comprehension of concept will be based on it. We also discuss the semi-automatic algorithm to adjust the concept-tree and the concept lexicon;(3)to design the classification and search algorithm based on concepts.