基于证据推理的信息不完全的多准则分类方法

A Multi-criteria Classification Approach Based on Evidential Reasoning for Incomplete Information

针对准则权系数信息不完全确定且有训练集的多准则分类决策问题,提出一种基于证据推理方法。该方法首先通过证据推理算法将准则值集成,得到集成效用值,在考虑训练集的基础上,结合不完全确定的准则权系数信息建立非线性规划模型。通过遗传算法求解该模型得到准则权系数和分类效用阈值,进而求出方案的效用值,并通过与每一分类的效用阈值比较,得到整个方案集的分类。最后实例说明该方法的有效性和可行性。

For the multi-criteria classification problems with values is incomplete and uncertain, a method based on evidential reasoning is proposed. Through evidential reasoning algorithms, the criteria values are aggregated and then the aggregated utilities are attained. Considering the training set, a nonlinear programming model is developed with the incomplete information on weights. Then the nonlinear programming model is solved by using genetic algorithms, and the optimal criteria＇s weights and utility threshold that defines the bounds of each category are gained. Then classification is performed through the comparison of the utilities of the alternatives to the utility threshold. Finally, an example is given to explain the feasibility and availability of this method.