Fig. 1. Energy level transition of the molecule

Fig. 2. Optical path of grating dual-beam infrared spectrometer

Fig. 3. Structure principle diagram of Fourier IR spectrometer

Fig. 4. Infrared spectrum database model

Fig. 5. Basic flow chart of PLS algorithm

Fig. 6. Schematic diagram of BP neural network

Fig. 7. RF schematic diagram of random forest model

Fig. 8. Schematic diagram of support vector machine

Fig. 9. Comparison of CDFOA-BP with four other models (left) and prediction of octane content in gasoline through RF (right)

Fig. 10. Scatter plot of PLS predicting methanol (left) and ethanol (right) content in gasoline

Fig. 11. Correlation between the predicted value of gasoline model No. 93# and the measured value of the classical method

Fig. 12. Correlation between the predicted and measured values of cetane content in diesel oil. (a) PLS model; (b) ERT model

Fig. 13. Scatter plots of PLS prediction of diesel methanol (left) and PLS-SVM prediction of ethanol (right) content