Table 2.

Recent studies performed for the diagnosis of clinical conditions using RQA-based ECG analysis.

Clinical conditions	Classification method	Performance	Ref.
Atrial fibrillation, atrial flutter, ventricular fibrillation, and normal sinus rhythm	Decision tree, random forest, and rotation forest	98.37%, 96.29%, and 94.14% accuracy for rotation forest, random forest, and decision tree, respectively	[123]
Effect of the exposure to low-frequency noise of different intensities on the cardiovascular activities	Statistical analysis of RQA-based measures	Statistically significant parameters obtained with p value ≤ 0.05	[126]
Obstructive sleep apnea	A soft decision fusion rule combining SVM and neural network	86.37% sensitivity, 83.47% specificity, and 85.26% accuracy	[128]
Arrhythmia	Joint probability density classifier	94.83 ± 0.37% accuracy	[129]
Sudden cardiac death	K-NN, SVM, decision tree, and probabilistic neural network	86.8% accuracy, 80% sensitivity, and 94.4% specificity with K-NN classifier and 86.8% accuracy, 85% sensitivity, and 88.8% specificity with PNN	[127]
Atrial fibrillation	Unthresholded recurrence plots	72% accuracy	[130]