Table 7.

The experimental performance of the proposed model (DNE-RL) vs other competitive deep learning benchmarks for Kaggle dataset.

Metric		GA	DE	PSO	GSK	MobileNet	VGGNet19	ResNet50	DenseNet12	DNE-RL
	AVG	0.933488	0.924888	0.942884	0.948021	0.949502	0.950001	0.972881	0.966913	0.991441
	STD	0.027901	0.012288	0.011142	0.014252	0.017881	0.023315	0.020994	0.013637	0.011042
ACC	Best	0.947991	0.937786	0.948788	0.958893	0.958812	0.961152	0.982441	0.971189	0.993668
	Worst	0.912298	0.915658	0.931229	0.932285	0.937881	0.927748	0.955571	0.952449	0.980671
	AVG	0.949893	0.940174	0.956618	0.957782	0.958838	0.968841	0.982441	0.973861	0.993568
	STD	0.037716	0.020781	0.030083	0.025151	0.023878	0.022274	0.027313	0.019669	0.021339
Precision	Best	0.968668	0.962881	0.972335	0.973991	0.968816	0.975561	0.986004	0.983315	0.996812
	Worst	0.908915	0.921007	0.913806	0.931401	0.927175	0.944767	0.952721	0.955672	0.968814
	AVG	0.925933	0.912829	0.935581	0.929181	0.935535	0.942552	0.960112	0.953443	0.981445
	STD	0.014047	0.018626	0.007996	0.009158	0.014663	0.018596	0.010052	0.008492	0.011381
Recall	Best	0.940116	0.936682	0.941887	0.937736	0.948778	0.955778	0.968784	0.959928	0.987886
	Worst	0.913005	0.904334	0.929981	0.920704	0.923051	0.932994	0.953008	0.942217	0.970221
	AVG	0.932717	0.921633	0.940081	0.945771	0.947006	0.948003	0.969951	0.964412	0.989666
	STD	0.015542	0.008797	0.012252	0.002363	0.006666	0.011882	0.016672	0.012331	0.004881
F-measure	Best	0.938812	0.935542	0.948788	0.949772	0.951332	0.954481	0.976866	0.969982	0.990454
	Worst	0.920094	0.916652	0.931176	0.938182	0.942553	0.935991	0.951331	0.957871	0.983353
	AVG	0.934451	0.926656	0.940092	0.949882	0.951666	0.947781	0.974554	0.963999	0.990337
	STD	0.022444	0.011434	0.008892	0.006772	0.021033	0.014403	0.009037	0.014544	0.007881
AUC	Best	0.945333	0.929922	0.948871	0.951991	0.969912	0.952662	0.979988	0.970002	0.992008
	Worst	0.911329	0.915004	0.934505	0.943303	0.936766	0.932242	0.968998	0.949889	0.982662