Table 2.

Statistics for the results of the RF model prediction at various lead-times.

Random forest model statistics
Prediction size (forecast)	Train RMSE [°C]	Test RMSE [°C]	Train r2	Test r2	Train F1	Test F1
6 h	1.56	1.72	0.98	0.98	0.96	0.95
12 h	1.86	2.07	0.98	0.97	0.94	0.94
24 h	2.12	2.37	0.97	0.96	0.94	0.94
30 h	2.45	2.83	0.96	0.94	0.93	0.93
36 h	2.6	3.01	0.95	0.94	0.92	0.92
48 h	2.86	3.16	0.94	0.93	0.91	0.92

Statistics shown include root mean square error (RMSE °C), correlation coefficient (r²), and the F1 score. The training statistics for RMSE and r² are the mean of 3 cross-validation folds.