食品安全网络舆情本体学习研究  

Research on ontology learning of Internet public sentiment concerning food safety

作　　者：李宏伟[1] 林萍[1] 洪小娟[1] 

机构地区：[1]南京邮电大学管理学院,江苏南京210023

出　　处：《南京邮电大学学报（社会科学版）》2013年第4期72-77,共6页Journal of Nanjing University of Posts and Telecommunications(Social Science Edition)

基　　金：教育部人文社科基金项目"移动互联网热点舆情与网下集合行为的耦合性研究"(11YJC630059);江苏省教育厅高校哲学社会科学研究基金资助项目"基于环境演化的食品供应链安全风险转移机制研究"(2012SJB630051);南京邮电大学校科研基金项目"基于Web2.0的高校隐性知识共享研究"(NY210056)

摘　　要：食品安全网络舆情本体在一定程度上可以提升舆情信息的抽取与检索效率,但海量的文本信息使得该领域本体的构建需要借助本体学习技术。本体学习的主要任务是概念获取以及概念间关系获取。研究中概念的获取采用了基于领域相关度和领域均匀度综合评价方法,概念间关系的获取采用了基于关联规则的概念间关系分析方法。考虑到中文词语的构成,在关联规则分析的基础上,进一步采用了基于词距判定的概念间关系分析方法,这可在一定程度上识别新概念及概念间关系。最后采集了229篇"立顿茶包"事件相关文档,采用上述方法构建了领域本体。The ontology of Internet public sentiment concerning food safety can promote to some extend the ex- tracting and searching efficiency of public sentiment information, but the ontology building on this issue needs to use the ontology learning technology due to its massive text information. The main tasks of ontology learning are concept extraction and concept relation extraction. In this research, the concept extraction method uses domain rel- evance and domain consensus, and the concept relation extraction method uses associative rules. Considering the composition of Chinese words, it further analyzes the concept relation based on judgment of distance between two words, which can be used for finding new concept and identifying some type of concept relation. In the end the do- main ontology is learned from 229 domain documents of Lipton tea bags event.

关 键 词：本体学习 食品安全 网络舆情 

分 类 号：G206[文化科学—传播学] TP182[自动化与计算机技术—控制理论与控制工程]