Optimizing Diabetes Risk Prediction: Metaheuristic- Driven Tuning of Deep Neural Networks

Diabetes encompasses a range of metabolic conditions triggered by consistently elevated blood sugar levels, demanding vigilant monitoring and tailored treatment. This disease predicted in its early stages can mitigate numerous health complications, averting potential harm to the kidneys, eyes, ulcers, heart, lungs, and blood vessels. In recent years, Deep learning advances have proven to be more effective than other methods in predicting diabetes. In this work we used Deep Neural Network (DNN) on Pima dataset for predicting Diabetes mellitus. We have leveraged a metaheuristic approach to meticulously refine parameters, achieving peak performance for the DNN which has resulted in 93.7% accuracy.