Table 4

Parameter tuning on the AD dataset

Parameters
Round	1	2	3	4	5	6	7	8	9	10*
Learning rate	0.0001	0.0001	0.001	0.001	0.001	0.001	0.001	0.0001	0.01	0.001	0.001
Batch size	32	16	16	8	32	16	8	8	8	8	16
Epochs	50	50	50	30	30	30	30	30	30	30	40
Dropout	–	0.3	0.3	0.15	0.15	0.15	0.15	0.15	0.15	0.15	0.2
Batch norm.	–	x	x	x	x	x	–	x	x	x	x
Metrics
Loss	1.040	0.711	0.637	0.461	0.742	0.600	2.292	0.805	0.639	0.600	0.677
Acc	0.833	0.794	0.802	0.840	0.833	0.840	0.728	0.767	0.825	0.833	0.837
Precision	0.881	0.977	0.891	0.768	0.947	0.949	0.628	0.972	0.788	0.859	0.948
Recall	0.628	0.447	0.521	0.809	0.574	0.596	0.628	0.372	0.713	0.649	0.585

AD, Alzheimer’s disease.