Machine Learning for Early Detection of Cardiovascular Disease in Fiji

Cardiovascular disease (CVD) has long been a leading cause of premature death in Fiji, accounting for approximately 80% of annual fatalities. The consequential burden of these diseases has been significantly aggravated by their late detection, thereby influencing the substantial rise in fatalities and casualties. Presently, Fiji lacks research in machine learning-based predictive models that could potentially detect early signs and novel patterns of CVD, creating a critical need for innovative approaches in the field. To bridge this gap, this study presents a pioneering method utilizing ensemble machine learning to enhance CVD diagnosis in Fiji. This approach incorporates the k-means algorithm and five supervised machine learning algorithms: K-nearest neighbor (KNN), Decision tree, Random Forest, Naïve Bayes, and Support vector machine [SVM]. Subsequently, the precision of these models was evaluated using 10-fold cross-validation, aiming to identify the most efficacious machine learning model for CVD prediction. Based on the accuracy tests, the Random Forest model emerged as the most accurate among the four models, achieving an average accuracy of 89% and demonstrating its suitability for predicting CVD in Fiji.