Machine Learning Approach to Predict Ventricular Tachycardia Using Heart Rate Variability Features

This paper comprehensively explores the classification of ventricular tachycardia (VT) patients and healthy individuals. The study employs Heart Rate Variability (HRV) analysis alongside machine learning techniques. HRV features are extracted from the Frequency domain using three distinct methods: Fast Fourier Transform (FFT), Autoregressive (AR) modeling, and Lomb-Scargle Periodogram (LSP). The dataset comprises 50 VT patients sourced from the Spontaneous Ven-tricular Tachyarrhythmia Database and 50 healthy subjects from the Normal Sinus Rhythm RR Interval Database, each with 12 hours of recorded data. Seven machine learning (ML) classifiers, including Logistic Regression, Decision Tree Classifier, Random Forest Classifier, Gaussian Naive Bayes, Support Vector Classifier (SVC), Multi-Layer Perceptron (MLP) Classifier, and k-Nearest Neighbors (KNN) Classifier, were employed for the classification task. The performance of these classifiers was rigorously assessed using a range of metrics, encompassing Accuracy, Sensitivity, Specificity, Precision, Negative Predicted Value (NPV), F1 Score, and Area Under the Receiver Operating Characteristic Curve (AUROC). The results highlight the efficacy of the FFT method and propose optimal classifiers for VT classification.