长波近红外光谱-偏最小二乘法快速分析白酒中乙醇含量

研究近红外光谱结合偏最小二乘法(NIRS-PLS)快速测定白酒中的乙醇含量的可行性,应用NIRS-PLS所建的模型相关系数达到0.99991,校正均方根误差(RMSEC)为0.00181,通过交互验证得出交互验证均方根误差(RMSECV)为0.00296,预测参差平方和(PRESS)为0.00016.用模型对预测集和白酒样品进行预测,预测均方根误差(RMSEP)为0.00258,结果表明NIRS-PLS可用于白酒生产中的在线质量监控和白酒市场的快速质量检测.

Rapid Determination of Alcohol Content in Distilled Spirits Using Long Wave NIRS With PLS Regression Method

This paper has studied the feasibility of quick determination of alcohol content in Distilled Spirits using long wave near infrared spectroscopy with partial least squares (NIRS-PLS) method. A calibration model is developed by NIRS-PLS. The regression coefficient (R) of this model is 0.99991. The root mean square error of calibration (RMSEC) is 0.00181. The root mean square error of prediction (RMSEP) is 0.00137. This model has been verified by cross-validation. The root mean square error of cross-validation (RMSECV) is 0.00296. The predicted residual error sum of squares (PRESS) is 0.00016. We predicted alcohol content in the validation and Distilled Spirits purchased from the market by this model. The results show that NIRS-PLS can reach the accuracy degree which is required in the production of Distilled Spirits.