化学计量学
偏最小二乘回归
线性判别分析
数学
决定系数
残余物
相关系数
统计
色谱法
化学
分析化学(期刊)
模式识别(心理学)
人工智能
计算机科学
算法
作者
Xuefen Sun,Huiling Li,Yuan Yang,Haimin Hua,Ying Guan,Chao Chen
标识
DOI:10.1016/j.saa.2020.119346
摘要
The aim of this study is to explore the feasibility of detection and quantification of two cheap adulterants (maltodextrin and starch) in Chinese functional food, hawthorn fruits powder (HFP), by using near infrared (NIR) spectroscopy coupled with chemometrics methods. The partial least squares discriminant analysis (PLS-DA) models were developed to discriminate the adulterated HFP from the authentic HFP, while the partial least squares regression (PLSR) models were employed to determine the contents of adulterants. In order to yield the best results, various spectra pretreatment methods and wavelength selection methods were carefully investigated. The models’ qualities were assessed by the self-consistency test, the independent test and the rigorous leave-one-out cross-validation test. The metrics for the PLS-DA discriminative model included error rate, true positive rate, true negative rate and F1 score, while the metrics for the PLSR quantitative model were determination coefficient, root mean square error and residual prediction deviation. Finally, very satisfying results were obtained, which indicate that our method is quite robust and applicable, and thus has great potential for rapid detection of adulteration in powder of many other herbal plants or functional foods.
科研通智能强力驱动
Strongly Powered by AbleSci AI