ARMA-Welch HRV Features: Predicting Ventricular Tachycardia with ML

This paper presents a comprehensive exploration of ventricular tachycardia (VT) classification using Heart Rate Variability (HRV) analysis and machine learning techniques. The study specifically delves into the analysis of HRV features extracted from the Frequency domain, utilizing both the Autoregressive Moving-Average (ARMA) method and the Welch method. The dataset encompasses 48 VT patients sourced from the Spontaneous Ventricular Tachyarrhythmia Database and 48 healthy subjects from the Normal Sinus Rhythm RR Interval Database obtained from PhysioNet. Seven machine learning (ML) algorithm, including Logistic Regression, Decision Tree Classifier, Random Forest Classifier, Gaussian Naive Bayes, Support Vector Machine (SVC), Multi-Layer Perceptron (MLP) Classifier, and k-nearest Neighbors (KNN) Classifier, were employed for the classification task. Performance evaluation metrics such as Accuracy, Sensitivity, Specificity, Precision, Negative Predicted Value (NPV), F1 Score, and Area Under the Receiver Operating Characteristic Curve (AU-ROC) were utilized to assess classifier performance. The outcomes reveal that the Welch method demonstrates superior performance across the tested classifiers. The results and subsequent discussions underscore the efficacy of the Welch method and provide insights into optimal classifiers for VT classification.