带有可信度因子的直觉模糊近似推理方法

An Approximate Reasoning Approach with Reliability Parameter based on Intuitionistic Fuzzy Set

针对基于相似度的直觉模糊近似推理问题,提出一种基于加权相似度量的直觉模糊推理方法。首先定义一种新的直觉模糊相似度度量公式,加入权重参数解决各维特征分配不均匀的问题,弥补了现有直觉模糊相似度量的缺陷。然后构建基于直觉模糊产生式规则的直觉模糊近似推理模型,加入可信度因子解决了随机性引起的信息不确定问题,同时给出模型的推理算法和计算步骤。最后通过实例验证了该方法的实用性和有效性,其在意图识别、目标识别等信息融合领域有良好的应用前景。

To the problem of Intuitionistic Fuzzy approximate reasoning based on similarity measure, a reasoning approach based on similarity measure with weighted parameter is presented. First, a new similarity measure formula is proposed. This method can solve the problem of every character with asymmetrical weight by taking account into the weighted parameter, and get over limitations of similarity measure methods between Intuitionistic Fuzzy Sets existing. Then, the Intuitionistic Fuzzy approximate reasoning model based on production rule is constructed. The problem of information uncertainty caused by randomicity is solved by adding reliability parameter. At the same time, the reasoning algorithm of the model and the calculating process are given. Finally, the practicability and the validity is checked by an instance. The approach has a good application in the information fusion filed such as intention recognition and target recognition.