血糖近红外光谱分析的Savitzky-Golay平滑模式与偏最小二乘法因子数的联合优选

Joint Optimization of Savitzky-Golay Smoothing Models and Partial Least Squares Factors for Near-infrared Spectroscopic Analysis of Serum Glucose

利用偏最小二乘法(PLS)和光谱Savitzky-Golay(SG)平滑方法,建立血清葡萄糖近红外光谱分析的优化模型。基于最优单波数模型的预测效果,提出划分校正集和验证集的一种新方法。采用10000～5300cm-1和4920～4160cm-1的组合波段,光谱经过SG平滑处理,利用PLS方法建立定标预测模型。将平滑点数扩充为5,7,…,87(奇数),多项式次数扩充为n=2,3,4,5,6,得到包含582个平滑模式的14个平滑系数表。对所有平滑模式和PLS因子数(1～40)分别建立PLS模型。按照预测效果进行优选,得到最优SG平滑模式为1阶导数平滑,3、4次多项式类型,SG平滑点数为53,最优PLS因子数为7,最优RMSEP达到0.376mmol/L。所采用的划分校正集和验证集的方法、SG平滑模式的扩充、SG平滑模式和PLS因子数的联合大范围筛选能够有效地应用于近红外光谱分析的模型优化。

The optimal model for the near-infrared spectroscopic analysis of serum glucose was established by partial least squares（ PLS） and Savitzky-Golay（ SG） smoothing method. Based on the prediction effect of the optimal single wave number model,a new dividing method for calibration set and prediction set was given. The calibration and prediction models were established by PLS method adopting the combination bands of 10000-5300 cm-1 and 4920-4160 cm-1 with Savitizky-Golay（ SG） smoothing. By extending the number of smoothing points to 5,7,…,87（ odd） and polynomial degree to 2,3,4,5,6,fourteen smooth coefficient tables including 582 smooth modes were calculated. All PLS models corresponding to all smooth modes and all PLS factors（ 1-40） were constructed. The optimal model was selected by the prediction effect. And the derivation order was 1,the polynomial degree was 3 or 4,the number of smoothing points was 53,the optimal factor was 7 and the optimal RMSEP reach 0. 376 mmol/L. The dividing method for calibration set and predic-tion set,the extending of SG smoothing modes,large-scale optimization combining SG smoothing modes and PLS factors can be effectively applied for the model optimization of near-infrared spectroscopic analysis.