Machine Learning for Automatic Classification of Web Service Interface Descriptions

Abstract. We argue that automatic classification of web service interface description into a predefined set of categories can considerably speed up the task of finding compatible web services. This can greatly improve to es-tablish automatic connection between networked systems. To carry out the classification, we show that techniques derived from automatic document classification can be successfully applied. We describe how the standard doc-ument classification techniques need to be adapted for our task, and finally present the results of a number of experiments in classification of web service interface descriptions.