基于信息抽取的匿名用户兴趣描述  

An Anonymous User Interest Profile Based on Information Extraction

作　　者：王波 姚敏[1] 

机构地区：[1]浙江大学计算机学院,浙江杭州310027

出　　处：《华南理工大学学报（自然科学版）》2004年第z1期117-120,共4页Journal of South China University of Technology(Natural Science Edition)

摘　　要：为准确地获取动态的用户兴趣信息,提出了一种基于信息抽取的用户兴趣自动获取方法.该方法首先根据电子商务网站中商品的领域知识构建领域Ontology,然后基于该领城Ontology自动抽取用户事务的语义文本信息,并将结构化的信息存入内容数据库,最后通过分析这些数据库中信息的特征值取得用户兴趣的描述文件.在汽车销售网站上的应用实例表明了该方法的有效性.In order to get exact information about user interest, a method based on information extraction is proposed to automatically get user interest. Firstly, the domain Ontology is constructed according to the domain knowledge of commodities sold in e-commerce websites. Next, semantic data are automatically extracted from unstructured documents on the basis of the domain Ontology. The structured data are then stored into the content databases. Finally, a user interest profile is obtained by analyzing the characteristic values of the information in the databases. An application instance of the proposed method in a website for car sale testifies its validity.

关 键 词：领域ONTOLOGY 信息抽取 语义分析 兴趣描述 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]