中文专利本体半自动构建系统设计

Design of the Ontology Semi-automatic Construction System Based On Chinese Patents

为充分发挥知识组织在企业专利战略中的作用,在分析专利文献的基础上,根据中文专利文献句法描述的特点,利用最大串频匹配、蚁群聚类、多层KMeans聚类、改进关联规则计算、基于规则和CRFs的术语关系抽取等算法,设计出一套领域本体的半自动构建系统,包括术语抽取、分类关系抽取、非分类关系抽取、本体形式化等模块,初步实现结构化数据和非结构化文本的本体半自动构建。

This paper builds a set of domain ontology semi -automatic construction system on the basis of patent documents analysis. This system contains such modules as the terminology extraction, the taxonomic relation extraction, the non-taxonomic relation extraction and the ontology formalization. And it achieves the ontology semi-automatic construction by structured data and non-structured texts preliminarily.