模糊推理的函数变换观点

基于函数论的立场 ,指出模糊推理过程是一个函数变换过程 ,模糊规则蕴涵了一个从函数空间到函数空间的映射 ,现存的种种模糊推理方法都是对这种映射的估计 ,进而指出插值和回归的方法都适用于这种估计。系统地提出了用回归的方法处理模糊推理的思想 ,并结合线性回归模型进行了示范 ,证明了基于线性回归模型的模糊推理系统 (FIS)同样是一个万能函数逼近器。

Functional transformation approach to fuzzy inference

Fuzzyinferenceisinessenceakind of functional mapping process.Fuzzy rules imply a mapping between functional spaces. Each of existing fuzzy inference methods is some estimation of such mappings. Both interpolation and regression approaches are applicable to this estimation. A systematic approach to deal with fuzzy inference is proposed and demonstrated by use of a linear regression model. Fuzzy inference systems (FIS) based on linear regression models are shown to be universal functional approximators.