A novel QoS prediction approach based on reversed and cross prediction

A novel quality of service(QoS) prediction approach DRaC based on the idea of collaborative filtering was proposed. This approach first employed clustering technology to handle the data-smoothing pre-processing, and then employed reversed prediction and cross prediction to handle the problem of data sparsity for the purpose of improving the prediction accuracy. Experiment results of QoS prediction demonstrate that the proposed DRaC approach outperforms existing methods in terms of accuracy. Moreover, the impact of each parameter to DRaC was evaluated.