摘要
介绍了径向基函数神经网络的原理、训练算法,建立的径向基函数神经网络城市需水量预测模型具有较强的非线性处理能力和逼近能力,运算速度快、性能稳定,克服了BP神经网络学习过程的收敛过分依赖于初值和可能出现局部收敛的缺陷,预测精度较高,泛化能力强。
The paper introduces radial basis function neural network theory,training algorithm,the establishment of radial basis function neural network model of urban water demand forecast has a strong nonlinear approximation ability and fast speed,stable performance,to overcome BP neural network learning process convergence and the possible over-reliance on the initial partial convergence of the defect,the higher prediction accuracy,generalization ability.
出处
《地下水》
2012年第1期114-116,共3页
Ground water
