Predicting Diabetes Using an MLP Neural Network Optimized by a Clonal Selection Algorithm

Diabetes, a chronic condition arising from the body's inability to produce or effectively utilize insulin, can lead to significant damage to vital organs such as the heart, blood vessels, eyes, kidneys, and nerves if left unchecked. Thus, early detection is crucial to mitigating these risks. In this study, we focus on predicting diabetes by employing a multilayer perceptron (MLP) neural network, which has been optimized using a clonal selection algorithm. Our model was tested on a dataset from a hospital in Frankfurt, Germany, containing detailed clinical and biological data of patients. This dataset empowers the model to deliver accurate and robust predictions regarding the likelihood of diabetes. The proposed approach shows great potential in enhancing diagnostic accuracy and tailoring treatment plans for individuals with diabetes.