Machine Learning Models to Predict Cardiovascular Events from Heart Rate Variability Data

Among the diseases known to mankind, cardiovascular diseases remain the deadliest and most expensive. However, Artificial Intelligence offers new solutions that can help diagnose these diseases and even predict their occurrence with high accuracy. In this study, we present several AI models that attempt to predict cardiovascular diseases. The models created are Support Vector Machines, AdaBoost, TabNet and TabTransformers. They were trained using Heart Rate Variability features extracted from the PhysioNet Smart Health for Assessing the Risk of Events via ECG Database. The models achieved high accuracies, which were 91.80%, 89.50%, 76.00% and 90.38% for the Support Vector Machines, AdaBoost, TabNet, and TabTransformers models, respectively.