Elimination of Irrelevancy during Semantic Service Discovery Using Clustering Approach

Semantic web service discovery process consumes a significant amount of time even to perform a typical service match of limited functional capabilities. Modern business integration requires services need to be dynamically discovered and composed quickly. In this work, a similarity based clustering approach is suggested for quick elimination of irrelevant services during discovery. In this method, all the available published services are clustered by service description similarity in prior to semantic matching. This yields clusters of similar services. After services are clustered, when a query is submitted, firstly, a particular cluster to which the query belongs to is found out. Then semantic matching of capabilities will be performed only to that particular cluster ignoring all other service clusters as irrelevant to the query. The proposed method is tested for its performance with a small test collection and results are presented.