Quantitative analysis of the oil mixture using PLS combined with spectroscopy detection

Partial Least Square (PLS) combined with ultraviolet (UV) spectroscopy and mid-infrared (MIR) spectroscopy were used to quantitative analysis of the oil mixture. The spectral data were obtained by ultraviolet (TU-1900) and infrared spectroscopy (IRTracer-100), and pretreated by the Savitaky-Golay (S-G) method. Then the regression models were established by the pretreated spectral data combined with PLS, and the predictive ability of the regression models were compared by the evaluation parameters (r(c), r(p), RMSECV, RMSEP and Error). The results show that: the evaluation parameters of the MIR-PLS regression model are: r(p) = 1, RMSECP = 0.0181, r(c) = 0.9929, RMSEV = 0.0111 and Error is 0.0361. The Error decreases 1.5% than that of UV-PLS regression model. It indicated that the MIR-PLS regression model has higher accuracy compared with UV-PLS regression model.

Automated summary

The study utilized Partial Least Square (PLS) combined with ultraviolet (UV) spectroscopy and mid-infrared (MIR) spectroscopy for quantitative analysis of oil mixture, with the MIR-PLS regression model demonstrating higher accuracy compared to UV-PLS regression model.