【摘要】 在对多元非线性回归模型和BP网络模型分析的基础上,将幂函数曲线回归模型和Baysiane规则化BP网络对滑动轴承合金材料的摩擦因数进行预测比较,分析了2种方法的预测效果不同的原因,并针对滑动轴承材料摩擦因数预测给出了以全样本进行训练的Baysiane规则化BP网络的结构。更多还原

【Abstract】 Based on the analysis of multi-element non-linear regression model and BP network model, the friction coefficient of sliding bearing materials was predicted by the power-law curve regression model and Baysiane regularized BP network. The reasons of their different results were analysed as well. The BP network structure trained by full-sample was proposed.更多还原