NB+: Diabetes Prediction using Novel NeuroBoost Algorithm NeuroBoost: A Novel Hybrid Deep Learning and Boosting Framework for Accurate Diabetes Prediction

Diabetes is among the major chronic diseases that have adversely impacted millions of individuals all around the world.Early prediction of diabetes is crucial as it will help the clinicians to provide timely medications and prevent health-related problems from happening.Past researches done on the topic of diabetes prediction utilized conventional machine learning techniques such as Logistic Regression, Random Forest, Support Vector Machines, and Artificial Neural Network (ANN).Though conventional machine learning models have proven to be effective in producing accurate results, they are still constrained with their inability to offer high levels of accuracy while predicting results and inability to establish the linkages between different health factors.In order to solve these problems, this research suggests the usage of NeuroBoost which is characterized by the uniqueness of the model since it is a combination of learning capability of neural network and boosting prediction capability.The main goal of the innovative model is to identify hidden patterns from data related to diabetes.The efficiency of the suggested approach was evaluated via testing on two widely used databases such as PIMA Indian Diabetes Dataset and real-world Primary dataset obtained from the Hospital.As a result of the experiments, the performance of the NeuroBoost model was identified as 98.00% accuracy with PIMA and 91% with primary dataset.Such results demonstrate the superiority of the suggested model in comparison with some other methods concerning the prediction of diabetes.Thus, the NeuroBoost innovative model can help healthcare practitioners identify patients who have high risks of developing diabetes and take necessary measures.