PLS回归方法在汽车滑行试验中的应用研究

A Study on PLS in Coasting Test of Vehicles

利用偏最小二乘法(PLS)对汽车滑行试验中的15个样本点数据进行建模分析,并将其与滑行阻力分析中常用的二次多项式拟合方法对比。结果表明,PLS回归模型能够较准确地预测汽车滑行阻力值,并有效地解决了样本点太少和变量间相关性严重的问题,PLS回归系数表明选取的自变量对汽车滑行阻力的影响趋势与实际情况是一致的。

Coasting test data of fifteen sample points were used to model and analyze coasting resistance by partial least-squares（PLS）regression method.Test results indicate that PLS model can make an accurate prediction,and solve some problems,such as very few test data are available and high correlations between variables exist.PLS regression coefficients show that the influencing trend acted on coasting resistance by the selected independent variables is similar to the actual situation.