Fusion of NIR and MIR Spectroscopy With Advanced CARS-PLS Techniques for Precise Quantification of Total Polyphenols...

The present study focuses on the application of near-infrared (NIR) spectroscopy, combined with mid-infrared (MIR) spectroscopy, to predict the levels of total polyphenols in peanut seed samples, highlighting innovative methodologies and advanced spectroscopic techniques.

To develop and validate accurate predictive models for quantifying total polyphenols in peanut seeds using NIR and MIR spectroscopy combined with advanced statistical approaches.

Partial least squares (PLS), competitive adaptive reweighted sampling PLS low-level (CARS-PLS Fusion_LL), and mid-level (CARS-PLS Fusion_ML) techniques were employed for model development. The total polyphenols were quantified using a spectrophotometer, and the model's efficacy was evaluated using calibration correlation coefficients (R_c), prediction correlation coefficients (R_p), root mean square error of cross-validation (RMSECV), root mean square error of prediction (RMSEP), and residual predictive deviation (RPD).

The CARS-PLS Fusion_ML Fusion method demonstrated high precision, with determination coefficients for prediction (R_p = 0.9818) and calibration (R_c = 0.9819). The RMSECV and RMSEP were calculated at 1.62 and 1.80, respectively, with an RPD value of 7.36.

In conclusion, a precise method for measuring the amounts of polyphenols in peanut seeds is provided by integrating NIR and MIR spectroscopy with advanced statistical methods. These findings underscore the potential of NIR and MIR spectroscopy, combined with advanced data fusion, for non-destructive, rapid, and cost-effective quantification of total polyphenols in peanuts, offering practical applications in the food industry for quality control and safety assessment.