本研究对多元线性回归模型中的残差部分再次进行回归分析 ,提取其中与自变量相关的成分 .用所得回归系数与初次回归所得系数相加作为新的回归系数 ,建立改进的回归方程 .将此建模方法用于近红外光谱定量分析 ,提高了预测精度 In this study, the residuals in the multiple linear regression model were regressed again to extract the components related to the independent variables, and the improved regression equation was established by adding the obtained regression coefficients and the initial regression coefficients as new regression coefficients. This modeling method is used for quantitative analysis of near-infrared spectroscopy and improves the prediction accuracy