采用贝叶斯正则化算法（抑制过拟合）提高 BP 网络的推广能力，采用两种训练方法， 即 L-M 优化算法（trainlm）和贝叶斯正则化算法（trainbr），用以训练 BP 网络；-Bayesian regularization algorithm (inhibition of over-fitting) to improve the generalization ability of BP network, using two training methods, that LM opti

BP网络实现分类问题；一，问题的提出；根据感知器的的相关理论易知感知器善于解决线性可分；反向传播网络（Back-PropagationN；一个具有r个输入和一个隐含层的神经网络模-The BP network realization categorizes a problem A, the problem puts forward Knows a machine according to the feeling of of the related theory easy to underst

模式分类中基于BP神经网络的曲线拟合；根据BP神经网络的模型，调用MATLAB工具箱函数，通过对样本数据进行训练，对测试样本数据进行成功的预测，预测的输出值与实际值之间几乎没有偏差。-Pattern classification of curve fitting based on BP neural network Based on BP neural network model, call MATLAB toolbox function, through the training sampl

基于BP神经网络的PID参数自整定，实现PID控制器的自适应控制(Self-tuning of PID parameters based on BP neural network)