Investigation of Chemometric Calibration Performance Based on Different Chemical Matrix and Signal-to-Noise Ratio

摘要

The chemometric calibration performance was systematically investigated by two parameters (changing the chemical matrix as well as the signal-to-noise ratio) of the NIR (near-infrared) spectrum. Three different analytes (hexane, cyclohexane, toluene) were selected and heptane was used as a solvent. The degree of spectral difference significantly affected the calibration performance. The largest structural difference between the analyte and the solvent provided the best calibration result for a given signal-to-noise ratio. Additionally, the signal-to-noise ratio of the spectra also directly influenced the calibration performance. Overall, the spectral difference and signal-to-noise ratio were the major factors for governing the chemometric calibration performance, especially in the low-concentration range.