Trust evaluation of CoPS partners based on wavelet support vector machine model and its application

Abstract

The integrators are facing the problems of so many decision attributes and few data samples for decision-making analysis when evaluating the partners of collaboration and innovation in complex products and systems(CoPS). Firstly, this paper created a trust evaluation model of collaborative partners in CoPS. Secondly, followed by the application of RS attribute reduction as a data pre-processing removes the redundancy in the decision-making property, and then combined with support vector machines (SVM) and wavelet function, this paper established wavelet support vector machines (WSVM) model, then used this model to accomplish the trust evaluation and classification of partners of collaboration and innovation. Compared with the traditional SVM, WSVM model can deal with the time series data, improve the classification performance, and reduce data dimensionality and the complexity classification of in the process, which helps decision-makers to achieve the confidence of partners in collaborative innovation evaluation and chosen. Finally, the method has been applied to the trust assessment process of CoPS System Integrators to core suppliers, and then paper elaborated the actual operation of the method steps and preliminary finished verification about the validity of the model.