以语义规则刻画服务请求的Top-k Web服务发现方法  

A Method of Top-k Web Service Discovery Using Semantic Rule Characterizing Service Request

作　　者：王海[1,2] 朱正东[1] 范琳[3] 李增智[1] 

机构地区：[1]西安交通大学计算机系统结构与网络研究所,西安710049 [2]北京邮电大学网络与交换技术国家重点实验室,北京100876 [3]西安邮电学院计算机系,西安710061

出　　处：《西安交通大学学报》2010年第2期35-38,共4页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(60673170,60803162)

摘　　要：为了更好地提供满足客户需求的Web服务,在充分研究服务发布描述、服务请求描述等特点的基础上,提出了以语义规则刻画的服务请求描述,并以本体构建了服务发布描述.通过对语义规则添加影响因子,使得服务匹配精度可以通过匹配度来度量,并按照给定的调节系数决定最终匹配是否成功.在OWL-S TC V2测试服务集合中进行了对比实验,结果表明,所提方法可以有效提高查全率和查准率,特别是在最大匹配数为20获得的k分别为2、5、8、10、15的情况下,Top-k查准率比基于描述逻辑包含检测的算法分别提高了15%、15%、9%、8%、2%,比关键字算法分别提高了5%、13%、30%、33%、30%,而查准率稳定在95%以上.In order to better meet customers＇ demands for the provision of web services, the semantic rules are used as the means to describe the service requests and the OWL-S service description language is used as service advertisement description language. The contribution of the semantic rules to the matchmaking results is reflected and the calculation of matching degree is facilitated by adding an impact factor to the rules in the service request description. An accommodation coefficient is introduced to judge if a given matching degree is enough to have a matchmaking success. Experiments using OWL-S TC V2 testing service set show that the proposed method increases the effective rate of recall and accuracy. Comparisons with the sub-sumption and the keyword matching based method in cases of k setting 2,5,8,10,15 and the maximum possible matching being 20 show that the top-k precision is promoted 15%, 15%, 9%, 8%, 2% and 5%, 13%, 30%,33%,30%, respectively, while the accurate rate remains stable at 95% or more.

关 键 词：WEB服务发现 语义规则 查准率 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]