Abstract
Diabetes is a prevalent and serious chronic illness that impacts millions of individuals globally. Early detection of diabetes is essential to mitigate severe health complications. This study investigates the application of Support Vector Machine (SVM) enhanced by Bayesian Optimization for the early prediction of diabetes. While SVM is a robust machine learning algorithm, its performance heavily depends on the proper selection of parameters. Bayesian Optimization is an efficient approach to fine-tune SVM parameters, such as the regularization parameter (C) and the kernel parameter (gamma). The research utilizes a Kaggle dataset that includes various diabetes risk factors. The study compares the performance of SVM optimized using Bayesian Optimization against SVM without optimization. The findings reveal that SVM with Bayesian Optimization achieves an accuracy of 95%, surpassing the 94% accuracy of the unoptimized SVM. These results highlight that Bayesian Optimization enhances SVM's effectiveness in predicting diabetes early
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call