偏最小二乘回归
化学
线性判别分析
校准
模式识别(心理学)
人工智能
定量分析(化学)
质量(理念)
生物系统
统计
色谱法
数学
计算机科学
哲学
认识论
生物
作者
Mengyin Tian,Ying Han,Xiaobo Ma,Wenyan Liang,Zhaoqing Meng,Guiyun Cao,Yi Luo,Hengchang Zang
摘要
Abstract Introduction Calculus bovis (C. bovis) is a typical traditional Chinese medicine (TCM) derived from animals, which has a remarkable curative effect and high price. Objectives Rapid identification of C. bovis from different types was realized based on spectral technology, and a rapid quantitative analysis method for the main quality control indicator bilirubin was established. Methods We conducted a supervised and unsupervised pattern recognition study on 44 batches of different types of C. bovis by five spectral pretreatment methods. Three variable selection methods were used to extract the essential information, and the partial least squares regression (PLSR) quantitative model of bilirubin by near‐infrared (NIR) spectroscopy was constructed. Results The partial least squares discriminant analysis (PLS‐DA) model could achieve 100% accuracy in identifying different types of C. bovis. The R 2 of the NIR quantitative model was 0.979, which is close to 1, and the root mean square error of calibration (RMSEC) was 2.3515, indicating the good prediction ability of the model. Conclusion The study was carried out to further improve the basic data of quality control of C. bovis and help the high‐quality development of TCM derived from animals.
科研通智能强力驱动
Strongly Powered by AbleSci AI