Optimized deep residual networks for early detection of myocardial infarction from ECG signals

Table 1 Comparative discussion

Datasets	Variations	Metrics	MSMIDM	SVM-GOA	CNN	kNN	Deep NN	ML-ResNet	hybrid ResNet-ViT model	Proposed SSS-DRN
*Dataset 1*	*k-value*	*Testing Accuracy*	0.775	0.787	0.808	0.829	0.854	0.878	0.883	0.901
		*Sensitivity*	0.784	0.796	0.816	0.836	0.859	0.879	0.887	0.905
		*Specificity*	0.786	0.798	0.819	0.841	0.866	0.882	0.889	0.907
	*Learning set*	*Testing Accuracy*	0.788	0.800	0.821	0.843	0.870	0.896	0.898	0.916
		*Sensitivity*	0.792	0.804	0.825	0.845	0.875	0.893	0.903	0.921
		*Specificity*	0.793	0.805	0.826	0.854	0.880	0.896	0.908	0.926
*Dataset 2*	*k-value*	*Testing Accuracy*	0.768	0.780	0.800	0.820	0.846	0.870	0.874	0.892
		*Sensitivity*	0.776	0.788	0.808	0.827	0.850	0.870	0.878	0.896
		*Specificity*	0.778	0.790	0.810	0.832	0.857	0.873	0.880	0.898
	*Learning set*	*Testing Accuracy*	0.780	0.792	0.813	0.835	0.861	0.887	0.889	0.907
		*Sensitivity*	0.784	0.796	0.817	0.837	0.867	0.884	0.894	0.912
		*Specificity*	0.785	0.797	0.818	0.846	0.871	0.887	0.898	0.917

ISSN: 1471-2261