Data Fusion Strategy Based on Near Infrared Spectra and Ultraviolet Spectra for Simultaneous Determination of Ginsenosides and Saccharides in Chinese Herbal Injection

Cheng Jiang,Yin Liu,Haibin Qu
DOI: https://doi.org/10.1039/c3ay26540d
IF: 3.1
2013-01-01
Analytical Methods
Abstract:A data fusion method based on near infrared (NIR) spectra and ultraviolet (UV) spectra for simultaneous determination of six ginsenosides and four saccharides in Chinese herbal injection (CHI) was developed. Two data fusion strategies (low-level data fusion and mid-level data fusion) combined with partial least squares (PLS) regression and uninformative variable elimination by PLS (UVEPLS) regression were implemented, respectively. Compared with the models established by independent NIR or UV spectra, there was a significant improvement provided by two data fusion strategies, which benefited from the synergistic effect of complementary information obtained from near infrared spectroscopy (NIRS) and ultraviolet spectroscopy (UVS). The results in this work showed data fusion of NIR and UV spectra combined with a regression algorithm could be a promising strategy to determine the ginsenosides and saccharides in CHI rapidly and simultaneously.
What problem does this paper attempt to address?