主成分-逐步回归近红外光谱法测量纸浆卡伯值

Pulp Kappa Number On-Line Determination by Near-Infrared Spectroscopy with Principal Component-Stepwise Regression Algorithm

主成分 -逐步回归法用于纸浆卡伯值的近红外光谱法测量建模。结果表明 ,这种算法由于主成分分析法消除了变量间的相关性 ,逐步回归法又在一定程度上防止主成分分析法带来的光谱数据信息的损失 ,所以利用该算法建立的纸浆卡伯值测量模型 ,与一元回归、多元回归和主成分回归等方法相比 ,具有更高的预测精度。

Principal component-stepwise regression(PCR-SRA) algorithm was proposed for th e determination of pulp kappa number by near-infrared spectroscopy.The results show that the precision of kappa number prediction is improved with PCR-SRA com pared with linear algorithms such as principal compo nent regression (PCR),multivariable linear regression (MLR) and unary linear reg ression (ULR).