Abstract
The two-layer fuzzy comprehensive evaluation (FCE) is a common evaluation method, but it is not applicable to the situation when factor sets of data sources have intersections. Due to this situation being common in today’s data deluge era, the two-layer FCE needs to be improved urgently. Therefore, this study proposes the multi-source FCE which does not limit whether factor sets have intersections or not. The objective of the multi-source FCE is to develop a comprehensive evaluation of an object (individual, product quality, customer credibility, etc.) on the basis of the evaluation data of each data source, whether factor sets of data sources have intersections or not. The underlying idea of the multi-source FCE is to fuse a multi-source FCE problem to an FCE problem. In the fusion process, this study forms an objective function to obtain optimal weights, and this makes the theoretical rationality of the multi-source FCE guaranteed. Finally, an example is given to illustrate the proposed method.
•The multi-source FCE is proposed to improve the two-layer FCE.•A fusion approach is put forward to solve the multi-source FCE problem.•An optimization function is formed to obtain the optimal weights.•The existence and uniqueness of the optimization function’s solution are discussed.•The gradient projection algorithm is used to solve the optimization function.