An Effective Web Service Selection based on Hybrid Collaborative Filtering and QoS-Trust Evaluation

Web Service mining has become one of the predominant areas of Service Oriented Architecture. Web service discovery methods include syntactic based system and semantic based system. In the proposed work, both syntactic and semantic based approach is followed. The most widely used recommender technique is collaborative filtering. In this paper, we proposed architecture for Web Service Selection based on Hybrid Collaborative Filtering (Memory and Model based Collaborative Filtering) and QoS-Trust evaluation. Memory based Collaborative Filtering is achieved by matrix formulation and Model based Collaborative Filtering is achieved by Cluster Based Hierarchical Probabilistic Latent Semantic Analysis – Feature extractor model. The proposed work addresses various collaborative filtering problems namely, cold-start problem, gray sheep problem, synonym problem, ramp-up problem, shillings attack, data sparsity and scalability. Keywords—Web Service, recommendation system, Collaborative Filtering, Cluster Based Hierarchical Probabilistic Latent Semantic Analysis – Feature extractor.