Big Data-Driven Machine Learning for Real-Time Heart Disease Prediction in Nigeria: A Case Study in a Lower-Middle-Income Healthcare Context

Abstract

Cardiovascular diseases (CVDs) are the leading cause of death globally, with a disproportionate impact on low- and middle-income countries like Nigeria. This study addresses the urgent need for effective early detection mechanisms by developing a real-time heart disease prediction system tailored to the Nigerian healthcare context. Leveraging big data from electronic health records (EHRs) and wearable devices, the research evaluates the performance of machine learning algorithms; Support Vector Machine (SVM), Logistic Regression, and K-Nearest Neighbors (KNN) in predicting heart disease. The dataset, sourced from Kaggle, underwent rigorous preprocessing, including handling missing values, outlier treatment, normalization, and feature selection. Model performance was assessed using metrics such as accuracy, Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Area under the Curve (AUC). Results indicate that SVM outperformed other models, achieving a test accuracy of 77.3%, MSE of 0.213, RMSE of 0.462, and AUC of 0.96, demonstrating strong generalization and predictive capabilities. Logistic Regression followed closely, while KNN exhibited overfitting and poor generalization. The study further developed a scalable prototype with user-friendly web and mobile interfaces to facilitate deployment in Nigerian health facilities. By integrating wearable data with EHRs, the system enhances early detection and real-time monitoring of heart disease, offering a viable solution to reduce the CVD burden in resource-constrained settings. This research contributes to the growing body of evidence supporting the application of machine learning in healthcare, particularly in LMICs, and underscores the potential of technology-enabled interventions in improving cardiovascular health outcomes.