An approach to identify candidate services for semantic web service discovery

A two-stage filtering approach based on implicit semantics of web service description and WordNet is proposed to identify candidate services during semantic service discovery using semantic reasoners. In the first stage, the candidate services relevant to the given service query are identified after filtering out irrelevant services. In the second stage, the trust score of each candidate service is compared against a user-defined trust score using Trust Rank algorithmic tool. Only those candidate services with trust score greater than the user-defined trust score are considered as the services which necessarily require complete semantic matching of service capabilities using a semantic reasoner. The proposed approach has been evaluated for accuracy using recall and precision. The results are compared with keyword based matching. With a typical test data of 90 services, the proposed method has average recall of 88.2% and average precision of 63.7% in contrast to keyword based matching which has average recall of 64.8% and average precision of 36.7%. Further, the proposed method has an average reduction of number of services of 97.3% and achieves optimization at matching level significantly.