Probabilistic approach for QoS-aware recommender system for trustworthy web service selection

摘要: We present a QoS-aware recommender approach based on probabilistic models to assist the selection of web services in open, distributed, and service-oriented environments. This allows consumers maintain trust model for each service provider they interact with, leading prediction most trustworthy consumer can with among plethora similar services. In this paper, we associate its performance denoted by QoS ratings instigated amalgamation various metrics. Since quality is contingent, which renders trustworthiness uncertain, adopt evaluation past experiences (ratings) consumers. represent using different statistical distributions, namely multinomial Dirichlet, generalized Beta-Liouville. leverage machine learning techniques compute probabilities belong classes. For instance, use Bayesian inference method estimate parameters aforementioned presents multidimensional embodiment corresponding also employ network classifier Beta-Liouville prior enable classification composite given constituents. extend our function an online setting Voting EM algorithm that enables estimation after interaction service. Our experimental results demonstrate effectiveness proposed approaches modeling, classifying incrementally ratings.