Table 4. Medium/Small metropolitan counties: Model performance comparison.

Medium/Small Metropolitan County Model
#	Models	Goodness-of-fit			Predictive accuracy
		R 2	RMSE	MAE	R 2	RMSE	MAE
1	Generalized Linear Model	0.626	0.398	0.300	0.570	0.402	0.307
2	Ridge Regression	0.626	0.398	0.300	0.570	0.402	0.308
3	Lasso Regression	0.591	0.416	0.312	0.537	0.418	0.317
4	Generalized Additive Model	0.779	0.305	0.233	0.645	0.364	0.274
5	Multi Adaptive Regression Splines [degree = 1]	0.750	0.325	0.249	0.655	0.359	0.272
6	Multi Adaptive Regression Splines [degree = 2]	0.760	0.319	0.246	0.627	0.371	0.280
7	Multi Adaptive Regression Splines [degree = 3]	0.790	0.297	0.230	0.587	0.391	0.287
8	Multi Adaptive Regression Splines [degree = 3; penalty = 2]	0.724	0.340	0.264	0.617	0.379	0.286
9	Random Forest	0.934	0.166	0.122	0.656	0.359	0.269
10	Gradient Boosting Method	0.967	0.117	0.090	0.620	0.376	0.284
11	Bayesian Additive Regression trees	0.804	0.287	0.218	0.667	0.354	0.266
12	Null Model (Mean-only)	NA	0.650	0.456	NA	0.619	0.440