论文部分内容阅读
目的开发多年份苹果糖度预测模型。方法将移动窗口偏最小二乘法(MWPLS)用于优化3个采收年份的苹果糖度信息区间,构建一种新颖的线性组合权重偏最小二乘法(LCW-PLS)模型。结果 MWPLS选择结果为4328~4787 cm-1、5323~5512 cm-1、5982~7135 cm-1和7159~7463 cm-1,当对应权重为0.004、0.070、0.066和0.860时,所建LCW-MWPLS模型预测性能较好,其RP=0.942、RMSEP=0.649%Brix和Q=0.890。结论 LCW-PLS法可改进常规PLS模型,为果品品质分级提供了一种建模参考方法。
Objective To develop a multi-year model of apple brix prediction. Methods A moving window partial least square method (MWPLS) was used to optimize the apple sugar content interval of three harvest years, and a novel LCW-PLS model was constructed. Results The results of MWPLS were 4328 ~ 4787 cm-1,5323 ~ 5512 cm-1,5982 ~ 7135 cm-1 and 7159-7463 cm-1. When the corresponding weights were 0.004,0.070,0.066 and 0.860, MWPLS model predicts better performance with RP = 0.942, RMSEP = 0.649% Brix and Q = 0.890. Conclusion The LCW-PLS method can improve the conventional PLS model, which provides a modeling reference method for fruit quality classification.