Real Time Prediction of Diabetes by using Artificial Intelligence

Abstract

The prognosis of diabetes, a chronic metabolic disorder, is crucial for early intervention and management. This study explores the integration of various machine learning approaches to enhance the accuracy and reliability of diabetes prognosis. The integration process encompasses rigorous cross-validation and hyper parameter tuning to optimize model efficacy. The findings demonstrate that leveraging multiple machine learning models not only increases prognostic accuracy but also provides a robust framework for handling diverse patient data. This integrated approach shows promise in supporting healthcare professionals with precise predictions, ultimately leading to better patient management and outcomes. The study also addresses the challenges of model complexity, computational cost, and interpretability, proposing solutions to mitigate these issues. The integration is done by stacking, an advanced ensembling method. The model developed is the results of accurately stacking the basic model. The divided data is examined using ML classifiers, and the classifier’s accuracy is calculated. Moreover, particular measures such as precision, recall, and F1 score are compared to each model’s prediction accuracy to choose the optimal model. The proposed method gives a high prediction accuracy of about 99 percentage.