基于分类优化的离散证据推理

离散信息在专家系统、模式识别、决策分析等领域普遍存在, 为了解决这类信息融合问题, 提出一种离散证据推理方法. 首先, 将每个离散证据拆分成一类单点值证据; 然后, 以冲突最小化为目标修正类内证据, 并采用证据推理进行组合; 最后, 以同样的方法对类间证据进行修正与组合. 所提出方法不仅可以解决离散证据的内外部冲突问题, 而且能够克服运算量过大的问题. 算例分析表明了所提出的方法是合理且有效的.

Discrete evidence based on grouping optimization

Discrete information widely exists in many areas such as expert systems, pattern classification, and decision analysis. To solve the problem on the kind of information fusion, a method for the combining evidence of discrete intervalvalued belief structures is proposed. Firstly, each discrete evidence is separated into a kind of single-valued evidences. Then, the bodies of intra-group evidences are modified with the objective of minimizing evidence conflict and combined by using the evidential reasoning(ER) approach. Finally, the bodies of extra-group evidences are modified and combined by using the same method. It not only solves the internal and external conflict problems of the discrete interval evidence, but also greatly reduces the computational complexity. The numerical examples show the efficiency and rationality of the proposed approach.