Deviation-based neighborhood model for context-aware QoS prediction of cloud and IoT services

Abstract How to obtain personalized quality of cloud/IoT services and assist users selecting appropriate service has grown up to be a hot issue with the explosion of services on the Internet. Collaborative QoS prediction is recently proposed addressing this issue by borrowing ideas from recommender systems. Going down this principle, we propose novel deviation-based neighborhood models for QoS prediction by taking advantages of crowd intelligence. Different from existing works, our models are under a two-tier formal framework which allows an efficient global optimization of the model parameters. The first component gives a baseline estimate for QoS prediction using deviations of the services and the users. The second component is founded on the principle of neighborhood-based collaborative filtering and contributes fine-grained adjustments of the predictions. Also, contextual information is used in the neighborhood component to strengthen the predicting ability of the proposed models. Experimental results, on a large-scale QoS-specific dataset, demonstrate that deviation-based neighborhood models can overcome existing difficulties of heuristic collaborative filtering methods and achieve superior performance than the state-of-the-art prediction methods. Also, the proposed models can naturally exploit location information to ensure more accurate prediction results.