摘要

With increasing number of enterprises involving in Cloud Manufacturing (CMfg), service reputation evaluation is becoming important in CMfg service platform. Existing reputation evaluation methods have poor configurablity and low scalability, they also suffer greatly from data sparsity. In this paper, we present a collaborative filtering (CF)-based hierarchical and configurable reputation evaluation method. First, we propose a hybrid framework to deal with sparsity issue in user feedback data. Then, we establish a library of evaluation metrics and select corresponding metrics to build a reputation computation model. Finally, the feedback data is fed into the evaluation model to calculate service reputations. We conduct an experiment to verify the validity and efficiency of our method.