摘要
A dynamic perceptron is designed by introducing time delay factors into the state functions of static perceptrons. The dynamic δ rule is deduced by using the gradient decent method. It can be used to identify a type of nonlinear process that the system′s inputs, its past inputs and states can influence the system’s outputs.
A dynamic perceptron is designed by introducing time delay factors into the state functions of static perceptrons. The dynamic δ rule is deduced by using the gradient decent method. It can be used to identify a type of nonlinear process that the system′s inputs, its past inputs and states can influence the system's outputs.
出处
《厦门大学学报(自然科学版)》
CAS
CSCD
北大核心
1999年第5期787-789,共3页
Journal of Xiamen University:Natural Science
基金
福建省自然科学基金