基于潜在语义分析的Web服务筛选技术  

Web Services Filtrate Technologies Based on Latent Semantic Analysis

作　　者：张孝国[1] 黄广君[1] 曹利红[1] 郭洪涛[1] 

机构地区：[1]河南科技大学电子信息工程学院,洛阳471003

出　　处：《计算机工程》2008年第15期39-41,共3页Computer Engineering

基　　金：教育部科学技术基金资助重点项目(03081)

摘　　要：Web服务匹配算法普遍缺少服务筛选机制,基于关键词对服务基本描述和服务质量描述进行匹配,会导致服务匹配效率低且准确率不高。该文提出一种基于潜在语义分析的Web服务筛选方法,将服务的基本描述和服务质量描述以树形结构属性模板表示,采用一定的词频统计和权重方法构建潜在语义空间,生成广告服务索引数据库,根据服务请求进行筛选。实验结果表明,该服务筛选方法具有较好的筛选准确率和筛选完全率,能够较大程度地提高服务匹配效率。The available services discovery mechanisms have lower matching efficiency and precision because of lacking services filtrate mechanism and matching the services basic attributes, quality of services attributes by keywords. This paper puts forward a Web services filtrate method based on Latent Semantic Analysis（LSA）. This method doscribes Web services basic attributes, quality attributes using tree structure, uses certain terms-frequency statistic method and weights method to build latent semantic analysis space, and builts advertising services index database and filtrates the Web services according to services request. Experimental results prove that this algorithm has higher precision and recall and improves services matching efficiency largely.

关 键 词：潜在语义分析 WEB服务 服务筛选 权重 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]