An Efficient Proposal for Deep Learning-Based Diabetes Prediction

As a result of its exponential growth in recent years, diabetes now poses a serious danger on a worldwide scale. As of right now, diabetes is one of the foremost causes of death worldwide. Failure to treat diabetes increases the risk of death from cardiovascular disease, cerebrovascular accident, brain stroke, and a host of other complications. In order to minimize further complications, it is crucial to discover diabetes early so that action can be taken to stop the condition from progressing. Predicting diabetes in its early stages may be possible with the use of cutting-edge machine learning techniques that uncover previously unseen patterns. Using the PIMA dataset and a variety of machine learning algorithms, this study lays forth an approach for diabetes prediction. All four of these functional classifiers—Deep Learning (DL), Artificial Neural Network (ANN), and Naive Bayes (NB)—achieve accuracy rates between 90% and 98%. With a 98.07% accuracy rate on the PIMA dataset, DL outperforms the other three methods when it comes to diabetes onset. In light of this, the suggested system offers healthcare officials a useful tool for prognosis. Incorporating omics data for illness onset prediction considerably improves the accuracy of the DL technique.