Table 3.

Accuracies and hyper-parameters of single dynamic learning networks.

Activation function	Accuracy and hyper-parameters	Mapping function type
		Linear	Tanh	Sinh
Case 1	Testing accuracy (%)	98.3333	97.2222	96.6667
	Training time (s)	6.0156	2.3906	1.4688
	Threshold ${\epsilon }^{\prime }$	0.2506	0.3270	0.2947
	Training rounds k	159	59	31
	NDLA coefficient $\alpha$	0.01	0.02	0.04
	Hidden neurons n	50	55	60
Case 2	Testing Accuracy (%)	98.3333	97.2222	97.7778
	Training time (s)	17.5000	2.7344	6.2344
	Threshold ${\epsilon }^{\prime }$	0.2154	0.2711	0.2935
	Training rounds k	290	40	109
	NDLA coefficient $\alpha$	0.01	0.04	0.01
	Hidden neurons n	42	40	41