论文部分内容阅读
采用近红外光谱技术结合化学计量学方法构建红曲米中红曲橙色素、红曲红色素、红曲黄色素的预测模型。分别采用多元线性回归(SMLR)、偏最小二乘回归(PLS)、主成分回归(PCR)构建所有色素组分的数学模型,以相关系数(R)、校正均方根误差(RMSEC)、预测均方根误差(RMSEP)、预测相对分析偏差(RPD)值来评价模型的综合性能。结果显示,MSC、SNV方法能够消除红曲米粉颗粒不均对光谱的散射影响;导数处理消除了基线漂移;对于红曲橙色素、红曲黄色素、红曲红色素三种模型均具有良好的稳定性;利用三种模型对未知红曲样品预测时,预测结果具有较高的线性,预测性能较好(RPD=2.86~5.39),可用于准确定量预测。结果表明近红外光谱技术可用于红曲色素的快速无损测定,为红曲米质量的智能化控制提供了新的途径。
A prediction model of red yeast orange, monascus red, and monascus yellow in red yeast rice was established by using near infrared spectroscopy and chemometric method. The mathematical models of all pigment components were constructed by using multiple linear regression (SMLR), partial least squares regression (PLS) and principal component analysis (PCR). Correlation coefficient (R), root mean square error of correction Root mean square error (RMSEP), predicted relative analytical bias (RPD) values to evaluate the overall performance of the model. The results showed that the MSC and SNV methods could eliminate the scattering effect of the uneven red yeast rice powder on the spectrum. The derivative treatment eliminated the baseline drift. For the three models of monascus orange, monascus yellow and monascus red, The predicted results are highly linear with good predictive performance (RPD = 2.86 ~ 5.39) when the three models are used to predict the unknown red yeast rice samples, which can be used for accurate quantitative prediction. The results show that NIR spectroscopy can be applied to the rapid nondestructive determination of monascus pigments, which provides a new way for the intelligent control of the quality of red yeast rice.