Heart Disease Prediction using an Ensemble Learning Method: A Study at King Abdullah Hospital in Bisha, Saudi Arabia

The detection of diseases is essential to improving healthcare outcomes and saving lives. Thanks to technological advancements in medicine, machine learning has become a valuable tool for predicting future patient health outcomes. Despite the abundance of available patient data, accurately predicting cardiac disease has become increasingly challenging. In response, we developed an innovative ensemble learning approach (ELA) that combines three powerful machine learning (ML) techniques. Our ELA provides reliable predictions of cardiac disease that surpass those of the individual classification algorithms, resulting in higher accuracy. Our research yields a new combination of classification algorithms that significantly increases the prediction accuracy. We tested our model on a regional dataset collected from King Abdullah Hospital in Bisha, Saudi Arabia. We obtained the best results false negatives (FN ) of 8, true positives (TP) of 70, true negatives (TN) of 72, false positives (FP) of 6, accuracy of 0.9113, sensitivity of 0.8839, specificity of 0.95, PPV of 0.9389, NPV of 0.8878, AUC of 0.9569, F1 of 0.9133 Kappa of 0.8220, MCC of 0.8277 with an ELA comprising logistic regression (LR), extra trees (ET) and support vector machine (SVM) with radial basis function (RBF) kernel. With our ELA, medical professionals can detect cardiac disease and provide timely interventions to prevent potentially life-threatening health issues.