基于多策略本体构建方法研究  

Research on Ontology Construction Method Based on Multi-strategy

作　　者：古凌岚[1] 孙素云[1] 

机构地区：[1]广东轻工职业技术学院,广州510300

出　　处：《计算机与数字工程》2011年第11期54-57,174,共5页Computer & Digital Engineering

基　　金：广东轻工职业技术学院自然科学研究项目(编号:KXKY200906)资助

摘　　要：针对手工构建本体工作量大、效率低以及更新维护困难等问题,文章提出了一种中文环境下多策略本体学习方法。使用统计分析和语义关联抽取术语,并利用构词模式发现分类关系,针对不同类型的非分类关系,分别采用句法模式、统计方法和基于规则方法,从而解决了现有本体学习方法对词典依赖性、处理中文效果差等问题。实验结果表明,该方法有较好地有效性和稳定性。Aim at the problem of ontology construction,which resulted in much work,low efficiency,and many difficulties in updating and maintenance ontology,a scheme of ontology learning with multi-strategy under the Chinese environment is proposed in the paper.Firstly,statistical analysis and syntax mode are used for extracting terms,association rules and syntax mode are used for finding taxonomic relationship,then,association rules and syntax mode are used for obtaining non-taxonomic relationship,accordingly,the problem of lexicon dependency and ineffective in traditional ontology learning methods is solved effectively.The experiment results showed that the proposed scheme provides better effectiveness and stability.

关 键 词：本体学习 术语抽取 关系抽取 多策略 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]