摘要
定义了代数神经元与代数神经网络，将符号计算融入代数神经网络，讨论多项式园式分解机理。设计出一类一元，多元多项式团式分解的神经网络模型。以整数城Z，复数域C上多项式为例，指出其网络模型的可行性。
The concepts of Algebra neurons and Algebra neural nfororks arefirstly proposed. We should symbol computation mix together algebraneurals networks and the factoirization ploynomials mechanism of algebraneural netwoks is discassed, and a kind of three lays forward algebraneural networks with single input and single oW and four lays forwardalgebra neural netwowks model of one or multivariate polynomials isdisigned, according to integer fields and complex fields of polynomialsand given factorization is practricable.
出处
《计算机工程与设计》
CSCD
北大核心
1999年第5期60-64,共5页
Computer Engineering and Design
关键词
神经网络
多项式
因式分解
近似符号计算
neural networks
algebra neural netwoks
polynomials networks
factorization
approximate symbol computaion
