Table 1. Evaluation of classical machine learning scores for all models, based on 5-features and 18-features inputs.

Model type	MSE	MSE	MAE	MAE	R2
	Train	Test	Train	Test
Constant prediction	3.71	3.72	1.36	1.31	0
Using 5 features:
LR + L1	2.91	2.91	1.11	1.11	0.21
LR + L2	2.92	2.93	1.12	1.12	0.21
LR + L1 + L2	2.86	2.87	1.11	1.11	0.23
GB-250	2.45	2.67	1.10	1.11	0.28
biLSTM RNN	2.36	2.90	0.92	1.01	0.33
Using 18 features:
LR + L1	2.77	2.77	1.09	1.09	0.25
LR + L2	2.69	2.69	1.08	1.08	0.27
LR + L1 + L2	2.67	2.68	1.07	1.07	0.28
GB-250	2.22	2.53	1.06	1.07	0.32
biLSTM RNN	2.03	2.45	0.85	0.90	0.43