三七
主成分分析
偏最小二乘回归
相关系数
光谱学
二维核磁共振波谱
分析化学(期刊)
生物系统
材料科学
化学
数学
人工智能
计算机科学
物理
核磁共振
色谱法
统计
医学
替代医学
病理
量子力学
生物
作者
Huo Zhang,Lanjuan Huang,Chuanpei Xu,Li Zhi,Xianhua Yin,Tao Chen,Yuee Wang,Guanglei Li
出处
期刊:Applied Optics
[The Optical Society]
日期:2023-06-29
卷期号:62 (19): 5306-5306
摘要
This paper proposes a Panax notoginseng (P. notoginseng) quantitative analysis based on terahertz time-domain spectroscopy and two-dimensional correlation spectroscopy (2DCOS). By imposing temperature perturbation combined with 2DCOS, the one-dimensional absorbance spectra were transformed into 2DCOS synchronous spectra, which reflected the differences in characteristic information between different P. notoginseng contents more clearly. Then, the feature information of P. notoginseng contents was extracted from the 2DCOS synchronous spectra by a competitive adaptive reweighted sampling (CARS) method and was used to build a quantitative model combined with a support vector regression machine (SVR), called 2DCOS-CARS-SVR. We obtained a more accurate analysis result than the commonly used principal component analysis (PCA)-partial least squares regression (PLSR) and PCA-SVR. The prediction set correlation coefficient and root mean square error reached 0.9915% and 0.8160%, respectively.
科研通智能强力驱动
Strongly Powered by AbleSci AI