Table 2.

Classification performance of future (IA) growth prediction models.

Model	Test data	AUC	accuracy	sensitivity	specificity
NSI	all	0.62	0.63	0.89	0.29
HMAX	all	0.52	0.54	0.36	0.80
AR	all	0.52	0.56	0.88	0.15
V	all	0.48	0.58	0.48	0.78
SA	all	0.48	0.56	0.36	0.84
RF	all	0.66	0.68	0.63	0.66
MLP	all	0.62	0.64	0.71	0.55
PointNet++ (dome)	all	0.72	0.77	0.80	0.63
PointNet++	all	0.795	0.82	0.96	0.63
RF	fold 1	0.56	0.55	0.33	0.80
RF	fold 2	0.81	0.81	0.83	0.8
RF	fold 3	0.63	0.64	0.66	0.6
RF	fold 4	0.63	0.72	1.0	0.4
MLP	fold 1	0.65	0.63	0.50	0.80
MLP	fold 2	0.61	0.63	0.83	0.4
MLP	fold 3	0.55	0.55	0.50	0.60
MLP	fold 4	0.67	0.72	1.00	0.4
PointNet++	fold 1	0.81	0.82	0.83	0.80
PointNet++	fold 2	0.70	0.73	1.0	0.4
PointNet++	fold 3	0.70	0.91	1.0	0.4
PointNet++	fold 4	0.75	0.82	1.0	0.5

The best result across all data is marked in bold.