Dynamic Ensemble Algorithm for Context-Aware Diabetes Prediction: Integrating Local and Global Data for Real-World Applications

This research proposes a novel decision support system utilizing dynamic ensemble algorithms for the accurate prediction of diabetes. This system mitigates the issues of regional disparities and unrecorded information in diabetes-related datasets by integrating global datasets and the medical data that are gathered locally. This technique utilizes the Adaptive Boosting (AdaBoost) algorithm blended with decision stumps and support vector machines to improve classification accuracy. The validation was carried out on the Pima Indian Diabetes Dataset and a locally curated dataset from multiple regions. The system that was suggested obtaining an accuracy of 84.5% was higher than the regular approaches. The article ends up with such suggestions as increasing the ensemble methods through the hybrid way.