摘要
采用中红外光谱获得羟丙基玉米淀粉中掺假原淀粉(玉米淀粉)的光谱数据,比较光谱数据的不同数学处理方法(求导阶数、平滑点数)对偏最小二乘法建立羟丙基玉米淀粉中掺假玉米淀粉预测模型的影响。当对光谱一阶求导和5点平滑时,模型校正集的决定系数(R_c^2)=0.984 0和均方根误差(RMSEC)=3.723%,羟丙基玉米淀粉中掺假水平实际值和所建立的模型预测值的相关系数为0.980 0。结果表明,中红外光谱法结合最小二乘法在羟丙基玉米淀粉掺假的快速鉴别上是可行的。
Mid-infrared spectroscopy was used to measure adulterated hydroxypropyl corn starch. Effect of mathe- matical processing methods, such as the number of derivation and the number of smoothing points, on the prediction model of adulterated hydroxypropyl corn starch built by partial least squares （PLS） was studied. The model （ using first derivative with 5 smoothing points） in the calibration set gave satisfactory performance with an R2o value of 0. 984 0 and root-mean-square error of 3. 723%. The correlation coefficient between the experimental values of various adul- teration levels and the predicated values calculated by the best predication model was 0. 980 0. The results indicated that it is possible to use mid-infrared spectroscopy combined with PLS in rapid identification of hydroxypropyl corn starch adulterated with corn starch.
出处
《食品与发酵工业》
CAS
CSCD
北大核心
2017年第4期216-221,共6页
Food and Fermentation Industries
基金
国家自然科学基金项目(31601430
31501446)
重庆市基础科学与前沿技术研究专项(cstc2015jcyjA80013)
西南大学第九届本科生科技创新基金立项项目(20161702006)
关键词
羟丙基淀粉
中红外光谱
掺假
偏最小二乘法
hydroxypropyl starch
mid-infrared spectroscopy
adulteration
partial least squares
