基于贝叶斯正则化BP神经网络的挤出温度预测模型

Extruding Temperature Prediction Model Based on Back Propagation Artificial Neural Network through Bayesian-regularization

简要介绍贝叶斯正则化BP神经网络原理,并应用基于贝叶斯正则化训练方法的BP神经网络建立挤出温度预测模型。预测与试验结果对比表明,经过训练后的网络模型基本获取了实际挤出温度的函数形式,网络输出值与样本对应的挤出温度实际值几乎完全重合,表明该方法能达到较好的预测精度,同时具有使用简洁、快速等优点。

The principle of back propagation（BP） artificial neural network based on Bayesian-regularization was briefly introduced,and the prediction model of extrusion temperature was established by using the BP artificial neural network.The comparison of prediction with test results showed that the trained network model correctively predicted the actual extrusion temperature function.This simulation method was very effective and simple.