Table 2 Model performance parameters for all six default models for both training and validation data averaged over 100 experiments

From: Dimensionally reduced machine learning model for predicting single component octanol–water partition coefficients

	\(\mathrm{RMSE}\)		\(\mathrm{MAE}\)		\({\mathrm{R}}^{2}\)
	Training	Validation	Training	Validation	Training	Validation
Linear	1.149	1.151	0.845	0.846	0.629	0.628
Ridge	1.149	1.151	0.846	0.847	0.629	0.628
Lasso	1.887	1.887	1.436	1.435	0.000	0.000
Random Forest	0.497	0.797	0.322	0.518	0.931	0.822
Gradient Boosted	0.961	0.988	0.702	0.718	0.741	0.726
Nearest Neighbors	0.743	0.905	0.502	0.616	0.845	0.770

Back to article page