基于文本挖掘的中文领域本体构建方法研究

Research on Construction methods of Chinese Domain Ontology Based on Text Mining

为了利用文本挖掘技术构建中文领域本体的方法,通过词性标注、依存句法分析、模式匹配等方法从非结构化的文本中自动抽取术语和关系。实验表明,本文的研究方法能够有效降低本体构建的复杂度,术语抽取更加全面,节省人力和时间成本。通过词形规则匹配和句法分析,自动抽取语义关系,包含整体部分关系、类属关系、关联关系以及用关联词表示的组合关系,语义更加丰富,构建的本体也能更好的反映领域的知识结构。

This paper proposes a construction method of Chinese domain ontology based on text mining technology. By using part-of-speech tagging, dependency parsing and pattern matching, we can automati- cally extract terms and relations from unstructured text. Experiment results show that the method can ef- fectively reduce the complexity of ontology construction, extract terms more comprehensive, save manpow- er and time cost. Through morphological rules matching and syntactic analysis,semantic relations can be automatically extracted which contains whole-part relations, generalizatin relations, association relations and syntagmatic relations represented by associated words.Semantics of the ontology is richer and the knowledge structure can be better reflected.