Publications

Abstract : With growing momentousness of Internet applications, digital world is overwhelmed with huge number of web services. To ease the job of selecting relevant WS in service composition process, recommendation system of Web Services is designed. It uses semantic analysis of WS along with enhanced collaborative filtering. Ontology based Semantic Analysis performed using Tversky Content Similarity Measure helps to identify most similar functionally relevant WS. The collaborative filtering process uses DBSCAN clustering and PCC similarity to identify highly collaborative WS, based on ratings given by experienced users. To overcome the existence of sparse data in WS ratings and to enhance filtering process, SVM Regression is implemented before collaborative filtering. Relative frequency method is applied to amalgamate collaborative and sematic similarity values of WS. The methodology is proved to produce more realistic, accurate and efficient WS recommendation. Future focus may be towards knowledge based filtering with real world contextual information.