利用近红外光谱和偏最小二乘回归法预测脂肪酸组成

Determination of Vegetable Oil Compositions by Near Infrared Spectroscopy and Partial Least Squares Regression

采集了30种植物油样品在10000~55 00 cm-1范围内的近红外透射光谱,将所有样品作为校正集,随机抽取10种样品作为预测集,以气相色谱方法测得植物油中主要成分油酸、亚油酸、棕榈酸、硬脂酸的含量为参考值,应用偏最小二乘回归法建立了基于近红外光谱的测定植物油主要成分含量的校正模型。四种成分校正模型的交叉验证误差均方根为0.281 1%~1.496 4%,预测误差均方根为1.080 8%~18.063 0%,校正集的预测值与实测值的相关系数均大于0.99,预测集中除了棕榈酸的预测值与实测值的相关系数为0.817 9,其余均大于0.9。

The near infrared transmission spectrums of 30 kinds of vegetable oil were measured in the frequency ranging from 10000cm-1 to 5500cm-1.All samples were used as a calibration set,and ten samples randomly selected were as a prediction set.The measured contents of main components（i.e.palmitic acid,stearic acid,oleic acid and linolic acid） of the vegetable oils by gas chromatographic were employed as a reference volume.A calibration model based on near-infrared spectroscopy determination of the main component contents was set-up by using the partial least-squares regression method.Results： The cross-validation root mean square error of the four components is 0.2811%~1.4964%,and the root-mean-square error is 1.0808%~18.0630%.The correlation coefficients of the predicted and measured values of the calibration set are over 0.99.The correlation coefficients of the predicted and measured values of the prediction set are over 0.9,except for palmitic acid with 0.8179.