基于Web语料的概念获取系统的研究与实现

Concept Extraction and Verification from Web Corpus

互联网网页中存在大量的专业知识。如何从这些资源中获取知识已经成为10多年来的一个重要的研究课题。概念和概念间的关系是知识的基本组成部分,因此如何获取并验证概念,成为从文本到知识的过程中的重要步骤。本文提出并实现了一种自动从Web语料中获取概念的方法,该方法利用了规则、统计、上下文信息等多种方法和信息。实验结果表明,该方法达到了较好的效果。

There is a large amount of knowledge on the Web pages. How to intelligently acquire knowledge from the massive information on Web pages has become a very important task. Concepts as well as inter-conceptual relations and inter-attribute relations of concepts are the main parts of knowledge. Therefore how toacquire and verify concepts is an important step in the knowledge acquisition. This paper proposes a hybrid approach to automatically extract concepts from large Web corpus. The hybrid approach makes use of rules, statistic, and context information to identify and verify concepts. The experiment shows very good performance of this method for extracting concepts.