ECG Classification for Detecting ECG Arrhythmia Empowered with Deep Learning Approaches

Table 1

Limitations of previous work.

Studies	Dataset	Method	Findings	Limitations
Strodthoff et al. [2]	PTB -XL	ResNet and inception	Predicted accuracy 89.8%	-No data augmentation
				-Less accurate
Wasimuddin et al. [5]	ECG-ID	CAD and machine learning	Predicted accuracy 98.5%	-Handcrafted
				-Small dataset
Elgendi and Menon [6]	SRAD	Database supervised ML algorithms	Predicted accuracy 75.02%	-No data augmentation
				-Handcrafted
Hsu et al. [14]	MIT-DB	AlexNet and ResNet	Predicted accuracy 94.4%	-No data augmentation
Acharya et al. [16]	PTB DB	CNN layers	Accuracy 93.5% with noise and 95.22% without noise	-Less accurate
				-Less number of classes
Gaddam et al. [37]	MIT-DB	Alex net	Predicted accuracy 95.6%	-Less accurate
				-No data augmentation
Reddy and khare [32]	UCI dataset	Rule-based fuzzy classifier and feature reduction	Predicted accuracy 76.51%	-Handcrafted
				-No augmentation
				-Less accurate
Poudel et al. [34]	KVASIR dataset	CAD and machine learning	F1-score of 0.88	-Handcrafted
				-No augmentation
				-Less number of classes
Siddique et al. [39]	Private	Fuzzy inference system, deep extreme machine learning, and ANN	87.05%, 92.45%, and 89.4%	-Handcrafted
				-No augmentation
				-Less accurate