Table 4 Performance comparison of different traditional deep learning models and their combinations with the proposed method (Accuracy, Precision, Recall, F1-Score, and AUC on four target domains).

Method	Target domain: MIT-BIH Sup arrhythmia					Target domain: MIT-BIH arrhythmia					Target domain: INCART					Target domain: SCD-Holter
Method	Acc	Precision	Recall	F1-Score	AUC	Acc	Precision	Recall	F1-Score	AUC	Acc	Precision	Recall	F1-Score	AUC	Acc	Precision	Recall	F1-Score	AUC
MLP	0.365	0.462	0.428	0.444	0.538	0.622	0.741	0.648	0.691	0.731	0.745	0.856	0.782	0.818	0.894	0.471	0.612	0.505	0.553	0.562
+Ours	0.402	0.495	0.462	0.478	0.617	0.653	0.766	0.671	0.716	0.786	0.781	0.874	0.803	0.837	0.911	0.502	0.641	0.523	0.579	0.637
LSTM	0.389	0.471	0.439	0.454	0.554	0.641	0.755	0.662	0.706	0.758	0.768	0.862	0.789	0.824	0.902	0.486	0.624	0.511	0.562	0.578
+Ours	0.418	0.502	0.471	0.486	0.629	0.672	0.781	0.689	0.734	0.811	0.803	0.883	0.815	0.848	0.920	0.516	0.651	0.537	0.590	0.646
1DCNN	0.374	0.468	0.433	0.450	0.541	0.636	0.749	0.657	0.701	0.748	0.752	0.859	0.785	0.820	0.898	0.479	0.618	0.508	0.559	0.567
+Ours	0.409	0.497	0.466	0.481	0.622	0.666	0.774	0.684	0.727	0.799	0.789	0.878	0.807	0.841	0.917	0.509	0.645	0.528	0.583	0.639

Quick links

Search