基于语义相似度与信息量的Web服务标签优化

Web Service Tags Optimization Based on Semantic Similarity and Quantity of Information

Web服务标签是用户向Web服务添加的描述其功能或属性的关键词,用来提高服务相似度计算的准确率进而改善服务发现、服务组合和服务聚类等.目前不准确甚至错误的无效标签比例较高,影响了服务相似度计算的准确率.为此,提出一个Web服务标签优化模型WS-TOM,分析了Web上已有WSDL文档,给出一种考虑到编程风格和命名规范的特征提取方法,用于Web服务相似度计算;然后综合标签与WSDL的语义相似度和标签的信息量来对标签进行排名,降低不准确标签的影响.实验结果及分析验证了WS-TOM模型能够提高Web服务相似度计算的准确率.

Web service tags are terms annotated by users to describe the functionality or attributes of Web services, which are used to improve the accuracy of Web service similarity computing and thus improve service discovery ,service composition and service cluste- ring, etc. However the present proportion of imprecise and even false tags is high and affects the accuracy of service similarity compu- ting. In this paper,we propose a Web service tag optimization model named WS-TOM. Firstly,based on the analysis of WSDL docu- ment already existed,a feature extraction approach for the calculation of the Web service similarity is suggested,considering both the programming style and the naming rules. Then tags are ranked by employing both semantic similarity and the information quantity to reduce the impact of inaccurate tags. Experimental results and analysis show that the proposed model can improve the accuracy of Web service similarity computing.