ECG Classification for Detecting ECG Arrhythmia Empowered with Deep Learning Approaches

Table 3

Training and validation ratio of proposed CAA-TL model

Proposed CAA-TL model training and validation (80%–20%)
AlexNet, SqueezeNet, ResNet50
Classes	Actual number of images	Training (80%)	Validation (20%)
F	3000	2400	600
N	3879	3103	776
Q	2500	2000	500
S	4000	3200	800
V	3500	2800	700
Total	16879	13503	3370