摘要
应用近红外光谱漫反射技术,采用偏最小二乘法和留一全交叉验证方法对燕麦和苦荞样品中膳食纤维的含量进行预测.对光谱波长和信号预处理方法进行了优化.剔除异常值后得到实验值和本方法预测值的相关系数达到0.9272.结果表明,采用近红外光谱漫反射方法能够对燕麦、苦荞其它食品中膳食纤维含量进行快速定量或半定量分析.
Dietary fiber in oat and buckwheat was determined by near infrared（NIR） diffuse reflection based on a partial least-squares（PLS） model which is cross-validated using leave-one-out（LOO） method.The spectral wavelength range and preprocessing method were optimized.After deletion of 9 outliers based on statistical criteria,correlation coefficient of 0.9272 between the predicted and experimental values was obtained.Results showed that NIR technique can be used for fast quantitative and semi-quantitative analysis of dietary fiber in oat and buckwheat as well as other food products.
出处
《首都师范大学学报（自然科学版）》
2011年第5期30-33,共4页
Journal of Capital Normal University：Natural Science Edition
基金
北京市属高等学校人才强教计划资助项目(PHR20100718)
关键词
近红外光谱法
膳食纤维
偏最小二乘法
留一法
near-infrared spectrometry
dietary fiber
partial least squares
leave-one-out.
