Enhancing Diabetes Prediction and Management through Machine Learning: A Comparative Study
K. Sudha, C Lakshmipriya, P. J. Beslin Pajila, E. Venitha, M. Anita, R. Sıva Subramanıan
Abstract
Diabetes, defined as high blood sugar levels, which represents an important risk to a person's health and which may have many negative consequences if left uncontrolled. Among these risks are heart-related diseases kidney problems, hypertension, visual impairment, & potential harm to various vital organs. Rapid action and an accurate prediction are necessary for successful diabetes care. This study aims to make early diabetes prediction and increase accuracy by using machine learning techniques. Machine learning (ML) approaches provide an efficient approach for diabetes prediction by using patient information to develop models. This work predicts the onset of diabetes using ML classification techniques on a diabetes dataset. Among the techniques used in this study include KNN, RF, NB, and LR. Selecting the model with the greatest accuracy and measure of how effectively it predicts diabetes is the purpose of the study. Our experimental results show Naive Bayes works better compare to other machine learning techniques and achieves very high accuracy. The findings of this study will significantly change the diagnosis and management of diabetes. Naive Bayes have shown that selecting the best correct model makes to more effective early treatments and customized healthcare strategies. This research suggests ongoing efforts to handle diabetes by highlighting the importance of early detection as an important in reducing the long-term impact of this prominent chronic condition.