基于本体的DeepWeb查询接口分类.pdf

, Deep Web ; , ?, 1976 , , , ; , ?, 1940 , , , . ЧDeep Web 徐和祥, 王述云, 胡运发 ( , 200433) E-mail: xiangh x01@ sohu. com ?`?: 目前对于分类问题, 主要工作集中在文本或Web 文档的分类研究,而很少有对 deep Web 查询接口的分类研究. deep Web 源包括查询接口和查询结果,大量的deep Web 源的存在,对它们查询接口的分类是通向deep Web 分类集成和检索的关键 步骤. 本分提出一种deep Web 本体分类方法,包括: 分类本体的概念模型和由此产生的deep Web 空间向量模型( VSM) .试验表 明, 这种分类方法具有良好的分类效果,平均准确率达到91. 6%,平均查全率达到 92. 4%. ? ? ?: Ч?; Deep Web; ; : T P181`````: A``````? ? щ ?: 1000- 1220( 2008) 10- 1889-04 Ontology- based Query Interfaces Classif ication of Deep Web XU He-xiang , WANG Shu- y un, HU Yun- fa ( De p ar tment of Comp uting and Inf ormat ion T echnol ogy , F udan Univ ersity , S hang hai 200433, China) Abstract: To dat e, in terms of the classificat ion, ex isting wo r ks mainly focus o n classifying text s o r Web document s, and there is little in the deep Web. M any deep Web so ur ces ar e st ructured by pr ov iding str uctur ed quer y inter faces and r esults. Cla ssify- ing query inter faces int o domains is one of the cr it ical steps towar d the integ rat ion o f hetero geneo us Web sources. In this pa- per , we pr esent an Onto log y- based quer y int erfaces classificatio n, w hich includes a categ or y Onto lo gy mo del and a nov el w eight calculat ion o ver Vecto r Space Mo del. The ex per imental result s show that we can g et a per formance w it h av erag e pr eci- sion 91. 6% and av era ge r ecall 92. 4%. Key words: onto lo gy ; deep w eb; quer y interface; classification 1`?`? Intenet б,o ?p , o ?p . [ 1] , Go og le, Yahoo 32%? deep Web ? ?, ?MSN , 11%?deep Web . deep Web o ?p ? Web , , , . Ber gman[2] ?Б, 2000 96000 ? ?Web ,2004 ?4 450000 ?[ 3] ., Goo glea Yaho o o ?p , .