指纹(计算)
偏最小二乘回归
均方误差
生物系统
化学
相似性(几何)
数学
模式识别(心理学)
均方根
人工智能
计算机科学
分析化学(期刊)
色谱法
统计
物理
量子力学
图像(数学)
生物
作者
Ying Pang,Xiang Li,Wenbo Zhong,Dandan Gong,Qian Chang,Zhenming Zhong,Ping Guo,Guoxiang Sun
标识
DOI:10.1016/j.microc.2022.107804
摘要
Liquorice, as a nutritional plant with a long histroy, is used in food and medical fields. In this study, we explored the quality consistent of liquorice from different locations by using multiple fingerprint profiles, including High Performance Liquid Chromatography (HPLC) and Fourier Transform Infrared Spectroscopy (FTIR). In addition, the average method of systematic quantified fingerprint method (AMSQFM) was presented to combine the chemical information of chromatography and spectral fingerprint. Macro qualitative similarity (Sm) and macro quantitative similarity (Pm) were used to assess the quality of 55 samples in terms of both qualitative and quantitative aspects. As a result, all the samples were well differentiated and classified into 6 grades. Additionally, partial least squares (PLS) model was used to evaluate the feasibility of quantizing glycyrrhizic acid (GLA) and liquiritin (LQT) by FT-IR spectral quantized fingerprint. The explanatory ability (R2Y) and prediction ability (Q2) of model were all greater than 0.9 and 0.5 respectively, the root mean square error of estimation (RMSEE) and root mean square error of prediction (RMSEP) were both less than 2.0. In the end, two characteristic parameters of electrochemistry (EC), oscillation life (tosc) and maximum amplitude (ΔEmax), were used to predict the reducibility of liquorice. Models were accurate and had good predictive abilities with the appropriate parameters, five active components in liquorice all exhibited certain reducing ability. In conclusion, this study provides a new idea for evaluating the quality and reduction ability of liquorice.
科研通智能强力驱动
Strongly Powered by AbleSci AI