Multi-Years Diabetes Prediction Using Machine Learning and General Check-Up Dataset

Diabetes mellitus (DM) is a severe chronic health condition affecting millions worldwide. Left untreated, it can lead to various debilitating and potentially life-threatening complications. Predicting DM cannot be overstated, as early detection and prevention can help reduce the risk of long-term health problems. Recent advances in machine learning have enabled the development of powerful predictive models that accurately diagnose diabetes. However, most current research in DM prediction is limited to predicting one year. This paper is a comprehensive study of multi-year DM prediction using machine learning and general check-up (GCU) as its dataset. This research conducted two main experiments to achieve this goal: finding the best machine-learning technique for DM prediction among nine standard techniques and analyzing its performance using a multi-year GCU dataset. The first experiment concluded that the decision tree is the best machine-learning technique for DM prediction. The second experimental results demonstrate that the decision tree yielded significantly high confidence when predicting up to four years ahead.