基于多输入单输出折线模糊神经网络的单调性分析

Analysis for monotonicity of polygonal fuzzy neural networks based on multi-input and single-output

依据折线模糊数及其扩展原理,在多输入单输出情形下针对折线模糊神经网络的单调性进行了理论分析.在转移函数满足非负递增时,证明了多输入单输出折线模糊神经网络中每个代表元都是单调递增的,进而在该网络构成一类函数的泛逼近器时,研究了这类函数的单调性问题.

According to the polygonal fuzzy numbers and its extension principle,the monotonicity of polygonal fuzzy neural networks based on multi-input and single-output is analyzed.And then,the representative element of polygonal fuzzy neural networks with multi-input and single-output is proved to be monotone increasing when the transfer functions are nonnegative increasing functions.Further,if the network is the universal approximator to the class of functions,the monotonicity of this function is researched.