Web Service Classification Based on Automatic Semantic Annotation and Ensemble Learning

With the development of Web Service Technology, the quantity of the web services published on the Internet is increasing rapidly. Recognizing each web service intelligently becomes the key of efficiently using Internet. And the first step of recognization is to classify the web services accurately. To classify a huge amount of web services becomes a difficulty job. Therefore, in order to support applications of web services more effectively, an automatic web service classification method is needed. In this paper, the common WSDL files are regarded as the study object. Since web service is described by WSDL, the traditional document classification method cannot be applied directly. In the paper, a new method is proposed which applies automatic web service semantic annotation and uses three classification method: Naive Bayes, SVM and REP Tree, furthermore ensemble learning is applied. According to the experiment done on 951 WSDL files and 19 categories, the accuracy was 87.39%.