TABLE 2.

Performance of 14 models.

Metrics	R 2	MAE	RMSE	Accuracy of the predicted concentration within ±30 (%)of the actual concentration
Models
Liner regression	0.42	0.29	0.37	61.5
LASSO regression	0.32	0.30	0.40	48.7
Ridge regression	0.42	0.29	0.37	61.5
Elastic Net regression	0.18	0.32	0.39	53.9
Bayesian Ridge regression	0.34	0.29	0.37	56.4
KNN	0.27	0.36	0.44	51.3
SVR	0.28	0.29	0.35	53.9
Random Forest	0.45	0.29	0.38	51.3
XGBoost	0.54	0.25	0.33	74.4
LightGBM	0.48	0.27	0.35	61.5
CatBoost	0.36	0.30	0.36	53.6
NGBoost	0.37	0.29	0.36	56.4
AdaBoost	0.36	0.29	0.36	50.0
GradientBoosting	0.40	0.32	0.37	64.1

Abbreviations: KNN, K-nearest neighbor; SVR, Support Vector Regression; MAE, Mean Absolute Error; RMSE, Root Mean Square Error.