一种基于Bootstrapping的本体学习方法  

An Ontology Learning Method Based on Bootstrapping

作　　者：张俊[1] 高志强[1] 徐惠[2] 蔡施彦[1] 戴云徽[3] 

机构地区：[1]东南大学计算机科学与工程学院,江苏南京210096 [2]东南大学软件学院,江苏南京210096 [3]南京理工大学经济管理学院,江苏南京210094

出　　处：《南京师范大学学报（工程技术版）》2008年第4期56-58,共3页Journal of Nanjing Normal University(Engineering and Technology Edition)

基　　金：国家科技计划(2006BAK10B02)资助项目

摘　　要：提出了一种基于自扩展的本体学习方法用于获取领域术语.该方法只需提供少量种子术语和一个未标注语料库作为输入,由种子术语开始学习抽取模式,再由学习到的模式发现新的术语,进一步由新发现的术语学习新的抽取模式,如此循环迭代.实验结果表明,该算法能够产生较高质量的领域术语集合和抽取模式集合,这样的集合可用于相关领域的信息抽取.This paper presents a bootstrapping-based approach for ontology learning which can be used in field words ac- quisition. This approach only requires a small set of seed terms and an unmarked corpus as its input, and it learns ex- traction patterns using seed terms, and detects new field terms using patterns learned before, and again using newly ac- quired field terms to discover novel extraction patterns, and iterate on. Experiments show that this approach produces a relatively high-quality domain-specific dictionary and a set of extraction patterns which can be consequently applied to information extraction in related domain.

关 键 词：信息抽取 本体学习 自扩展 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]