文献类型：期刊文献

中文题名：Fourier三角基神经元网络的权值直接确定法

英文题名：A Direct-Weight-Determination Method for Trigonometrically-Activated Fourier Neural Networks

作者：张雨浓[1];旷章辉[1];肖秀春[1,2];陈柏桃[3]

机构：[1]中山大学信息科学与技术学院,广东广州510275;[2]广东海洋大学信息学院,广东湛江524088;[3]广东海洋大学航海教育质管办,广东湛江524088

年份：2009

卷号：31

期号：5

起止页码：112

中文期刊名：计算机工程与科学

外文期刊名：Computer Engineering & Science

收录：CSTPCD、、CSCD2011_2012、北大核心2008、北大核心、CSCD

基金：国家自然科学基金资助项目(60643004;60775050);中山大学科研启动费;后备重点课题资助项目

语种：中文

中文关键词：三角正交基函数;Fourier三角基神经元网络;权值修正;直接确定法

外文关键词：trigonometric activation function; trigonometrically-activated Fourier neural network; weight-updating formula ; direct determination method

中文摘要：根据Fourier变换理论,本文构造出一类基于三角正交基的前向神经网络模型。该模型由输入层、隐层、输出层构成,其输入层和输出层采用线性激励函数,以一组三角正交基为其隐层神经元的激励函数。依据误差回传算法(即BP算法),推导了权值修正的迭代公式。针对BP迭代法收敛速度慢、逼近目标函数精度较低的缺点,进一步提出基于伪逆的权值直接确定法,该方法避免了权值反复迭代的冗长过程。仿真和预测结果表明,该方法比传统的BP迭代法具有更快的计算速度和更高的仿真与测试精度。

外文摘要：Based on the Fourier transformation theory, a feed-forward neural network using trigonometric orthogonal activation-functions is constructed in this paper. The neural network adopts a three-layer structure, where the input and output layers employ linear activation functions, while the hidden-layer neurons are activated by a series of trigonometric orthogonal functions. In this paper, we first derive its weight-updating formula by adopting the standard BP training algorithrn. More importantly, a pseudo-inverse method is proposed as well, which directly determines the weights of the neural network without iterative BP training. Simulation results show that the direct-weight-determination method is more efficient and accurate than the conventional BP iterative-training algorithms.