Table 1.

Model performance comparison.

Model	$R^2$	In-sample		Out-of-sample
		RMSE	MAE	RMSE	MAE
Generalized linear model	0.987	0.387	0.295	0.389	0.295
Ridge regression	0.971	0.584	0.457	0.588	0.459
Lasso regression	0.962	0.661	0.517	0.665	0.519
Generalized additive model	0.991	0.315	0.238	0.319	0.240
Multivariate adaptive regression splines [MARS]	0.983	0.441	0.328	0.442	0.328
MARS [degree=2]	0.982	0.454	0.343	0.456	0.344
MARS [degree=3]	0.982	0.454	0.342	0.456	0.343
MARS [degree=3; penalty=2]	0.981	0.454	0.361	0.456	0.343
Random forest	0.996	0.199	0.139	0.493	0.347
Gradient boosting method	0.994	0.261	0.197	0.309	0.282
Bayesian additive regression trees	0.997	0.182	0.136	0.221	0.159
Null	NA	3.382	2.773	3.386	2.774