【作者】 潘涛； 韩筠； 彭爱红； 尹浩； 田佩玲； 韦相才；

【Author】 PAN Tao~1,HAN Yun~1,PENG Ai-hong~1,YIN Hao~1,TIAN Pei-ling~2,WEI Xiang-cai~(2*) 1.Key Laboratory of Optoelectronic Information and Sensing Technologies of Guangdong Higher Educational Institutes,Jinan University,Guangzhou 510632,China 2.Family Planning Research Institute of Guangdong,Guangzhou 510600,China

【机构】 光电信息与传感技术广东普通高校重点实验室(暨南大学)； 广东省计划生育科学技术研究所；

【摘要】 采用傅里叶变换红外光谱（FTIR）结合衰减全反射（ATR）技术、Savitzky-Golay（SG）光谱平滑方法和偏最小二乘（PLS）回归建立人体血红蛋白（HGB）的快速定量分析模型。利用计算机算法平台,将483种SG平滑模式和不同PLS因子数（从1～30）任意组合,分别建立SG平滑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,2次多项式,59平滑点数,最优PLS因子数为8,最优RMSEP,R_P和RRMSEP分别为2.11 g·L^-1,0.962和3.11%。结果表明,SG平滑模式和PLS因子数的联合优选可以明显改善模型的预测效果。更多还原

【Abstract】 The rapid quantitative analysis models of the human blood hemoglobin（HGB） were developed by FTIR spectrometer. ATR techniques,Savitzky-Golay（SG） smoothing method and PLS regression.By computer algorithm,for any combination of 483 SG smoothing modes and PLS factors（1-30）,PLS models were established respectively.By the prediction effect,the optimal SG smoothing mode was first order derivation,2 degree polynomial,59 smoothing points,the optimal PLS factor,RMSEP. Rp and RRMSEP were 8,2.11 g·L^-1,0.962 and 3.11%respectively.The results showed joint optimization of SG smoothing modes and PLS factors improved the model prediction effect obviously.更多还原