Diabetes Classification and Prediction Through Integrated SVM-GA

Diabetes is a global health concern. If the sources are to be believed, about 422 million people across the globe are suffering from this disease. It's when the body can't regulate sugar because it lacks insulin or can't use it correctly. Due to its chronic nature, it has become very important to identify it in the initial stage so that it can be treated at the right time. Machine learning (ML) has emerged as a valuable tool for diabetes prediction, and many researchers have employed ML techniques for accurate predictions. Among these, the support vector machine (SVM) is the most popular and widely used algorithm among researchers for predicting diabetes at an early stage. In this paper, researchers have proposed an integrated SVM-genetic algorithm (GA) method and evaluated its performance against various ML methods, such as SVM, random forest, K-nearest neighbors (KNN), decision tree (DT), extra trees classifier (ETC), Naive Bayes (NB), XG boost, and gradient boosting. And we employed the PIMA Indian Diabetes Datasets (PIDD) for this comparative analysis. The results of this study reveal that the proposed integrated SVM-GA method outperforms other methods, particularly in terms of accuracy, precision, recall, and area under the curve (AUC).