利用RBF神经网络实现高斯型函数积分

Implementation for Gauss-Type Function Integral Using RBF Neural Networks

导出了在一定精度下高斯型函数积分近似表达式,利用径向基函数(RBF)网络具有良好的逼近任意非线性映射的特点,提出了一种改进的RBF网络方法以实现对高斯型函数积分。实验结果表明所提出方法具有较高的逼近计算精度。

First an proximate expression of Gauss type function integral is deduced with proper accuracy, and then a scheme based on modified radial basis function (RBF) neural networks is proposed. The numerical experiments indicate that the proposed scheme has a higher proximate accuracy.