Table 9.

Top-1 combinations for the DenseNet169 in the 15 optimization iterations.

#	Parameters optimizer	Batch size	Dropout ratio	TL learn ratio	Loss	Accuracy	F1	Precision	Recall	AUC	WSM
1	AdaMax	32	0.37	50%	1.1045	78.11%	78.10%	78.11%	78.11%	0.8572	76.20%
2	RMSProp	64	0.5	70%	15.632	82.06%	82.06%	82.06%	82.06%	0.8228	80.01%
3	AdaMax	32	0.36	50%	1.7170	85.67%	85.66%	85.67%	85.67%	0.8845	83.55%
4	AdaMax	32	0.36	50%	1.9082	88.02%	88.02%	88.02%	88.02%	0.9028	85.85%
5	AdaMax	32	0.36	50%	1.6761	88.89%	88.89%	88.89%	88.89%	0.9100	86.70%
6	AdaMax	32	0.36	50%	1.7754	88.02%	88.02%	88.02%	88.02%	0.9028	85.85%
7	AdaMax	32	0.36	50%	1.1883	89.71%	89.70%	89.71%	89.71%	0.9261	87.50%
8	AdaMax	32	0.36	50%	1.2521	88.07%	88.06%	88.07%	88.07%	0.9117	85.91%
9	AdaMax	32	0.36	50%	1.7210	88.84%	88.84%	88.84%	88.84%	0.9103	86.65%
10	AdaMax	32	0.36	50%	1.1751	88.12%	88.11%	88.12%	88.12%	0.9153	85.95%
11	AdaMax	32	0.36	50%	1.0352	89.56%	89.56%	89.56%	89.56%	0.9290	87.37%
12	AdaMax	32	0.36	50%	1.6333	87.73%	87.73%	87.73%	87.73%	0.9005	85.57%
13	AdaMax	32	0.36	50%	1.3847	87.01%	87.01%	87.01%	87.01%	0.9043	84.87%
14	AdaMax	32	0.36	50%	1.1712	90.76%	90.76%	90.76%	90.76%	0.9294	88.54%
15	AdaMax	32	0.36	50%	1.0113	89.66%	89.66%	89.66%	89.66%	0.9244	87.47%