Table 4.

Accuracy values (r2) of ensemble DNN compared with other ML models to predict number of attempts

	Best attempt	Dataset size	Meta-leaner (ensemble DNN)	Random forests	Gradient boosted regression trees	Kernel SVM
PB2	5	95	0.87	0.86	0.83	0.69
RW2	3	140	0.89	0.88	0.89	0.83
MB1	3	103	0.91	0.89	0.89	0.87
RR2	4	119	0.89	0.87	0.83	0.72
TR2	2	165	0.96	0.94	0.94	0.88