Table 2:

Performance values (RMSE and R²) obtained for a random forest model trained with UV-visible spectrophotometry data (400–500 ηm), applying several pre-processing methods to the data.

UV-visible (400–500 ηm) + Preprocessing, random forest
	TCC Spectrophotometry
	RMSE	R 2
Smoothing interpolation	5.773	0.6053
Background and Offset corrections	5.936	0.5927
Background correction	6.175	0.5956
No preprocessing	6.194	0.5581
Scaling	6.447	0.5740
Background, Baseline and Offset corrections	9.397	0.4780
First derivative	10.774	0.4482
Multiplicative Scatter Correction	11.621	0.3245

The total carotenoid content (TCC) determined by HPLC was used as response prediction variable. The best performance values are represented in bold.