论文部分内容阅读
针对四类24种钢材,将其化学成分和t8/5性能的数据作为样本,构造学习空间,建立了一种新的预测方法BP网络预测模型,并讨论了网络参数对网络收敛速度及误差函数的影响
Aiming at four kinds of 24 types of steel, taking the chemical composition and t8 / 5 performance data as samples, a new prediction model of BP network was established by constructing learning space. The network parameters’ convergence speed and error function Impact