EPJ Data Science

Table 2 Model performance of random forests (RFs), multiple linear regression (cubic), and linear spline regression (spline)

From: Temperature impact on the economic growth effect: method development and model performance evaluation with subnational data in China

	80% training data model result				20% OOB test data for prediction
	\(R^{2}\)	RMSE	MAE	AIC	\(R^{2}\)	RMSE	MAE
RFs	0.93	0.04	0.03	2.78	0.31	0.12	0.09
cubic	0.35	0.11	0.09	14.12	0.04	0.19	0.16
spline	0.38	0.11	0.09	13.40	0.02	0.19	0.16
RFs_lag2	0.93	0.04	0.03	2.11	0.44	0.11	0.08
cubic_lag2	0.38	0.11	0.09	13.40	0.01	0.20	0.16
spline_lag2	0.42	0.11	0.09	12.59	0.00	0.21	0.16

Back to article page