一类拟插值Kantorovich型神经网络算子的估计  

Estimates for the quasi-interpolation neural network operators of kantorovich type

作　　者：项承昊 赵易 XIANG Chenghao;ZHAO Yi(School of Mathematics,Hangzhou Normal University,Hangzhou 311121,China)

机构地区：[1]杭州师范大学数学学院,杭州311121

出　　处：《华中师范大学学报（自然科学版）》2023年第2期195-200,共6页Journal of Central China Normal University：Natural Sciences

基　　金：国家自然科学基金项目(11601110)。

摘　　要：该文在神经网络算子理论中的Max-product型算子和Kantorovich型算子的基础上,构造了一种由Sigmiodal函数激发的拟插值型的神经网络算子,考虑了其对实数域上非负连续函数的点态逼近和一致逼近,并给出了其在L^(p)_(+)(ℝ)空间上的逼近定理.Based on the Max-product operators and Kantorovich type operators in the neural networks operators theory,a quasi-interpolation type neural network operator activated by Sigmoidal functions is constructed.The pointwise and uniform approximations of nonnegative continuous functions space are considered.Moreover,the quantitative estimate similar to K-functional on the L^(p)_(+)(ℝ)spaces is given.

关 键 词：神经网络算子 KANTOROVICH型算子 Max-product型算子 逼近 

分 类 号：O174.41[理学—数学]