A Review on the Nonlinear Dynamical System Analysis of Electrocardiogram Signal

Table 1

Recent studies performed for arrhythmia detection using phase space analysis of ECG.

Types of arrhythmia	Classification method	Performance	Ref.
Atrial fibrillation, ventricular tachycardia, and ventricular fibrillation	Distribution of the attractor in the reconstructed phase space	85.7–100% sensitivity, 86.7–100% specificity, and 95.55% overall efficiency	[12]
Ventricular tachycardia, ventricular fibrillation, and ventricular tachycardia followed by ventricular fibrillation	Box-counting in phase space diagrams	96.88% sensitivity, 100% specificity, and 98.44% accuracy	[89]
Ventricular fibrillation and normal sinus rhythm	Neural network with weighted fuzzy membership functions	79.12% sensitivity, 89.58% specificity, and 87.51% accuracy	[90]
Atrial premature contraction, premature ventricular contraction, normal sinus rhythm, left bundle branch block, and right bundle branch block	K-nearest neighbour	99.42% sensitivity, 98.19% specificity, and 98.7% accuracy	[88]
Soon-terminating atrial fibrillation and immediately terminating atrial fibrillation	A genetic algorithm in combination with SVM	100% sensitivity, 100% specificity, and 100% accuracy	[91]