Abstract: While building service oriented Web application, it is necessary to search domain related Web services. Current search in UDDI servers is based on taxonomy of tModel and keyword-matching, which is not convenient to find domain related Web services. By crawling WSDL files directly from the Internet, and utilizing support vector machine to build WSDL file classifier, a method is proposed to realize domain based Web services search. Experiment shows that the method has a good accuracy.

Key words: Web service search, Domain, Support vector machine, WSDL

摘要： 在构建面向服务的Web应用时，往往需要按照领域进行Web服务查找。现有的UDDI的Web服务查找方式是通过tModel分类信息和关键字匹配来进行的，不便于按照领域进行查找。通过直接在互联网上搜索WSDL文件，并利用支持向量机来构建基于领域的WSDL文件分类器，实现了按照领域进行Web服务查找的方法。实验证明该方法具有较高的精确度。

关键词: Web服务查找, 领域, 支持向量机, WSDL

CLC Number: 

• TP391