Unlocking The Potential Of Machine Learning For Diabetes Prediction

Abstract

Millions of individuals throughout the world suffer with diabetes, a chronic condition that if unchecked can have catastrophic health repercussions. In order to forecast diabetes risk and aid healthcare professionals in managing or preventing the condition, machine learning algorithms have become increasingly effective. The goal of our work is to inspect the achievement of machine learning techniques in predicting diabetes. The dataset used in previous study consists of demographic and clinical data of patients who have been diagnosed with diabetes and those who have not. Different classification and Neural Network algorithms, such logistic regression, Artificial Neural Network, XGBoost Random Forest, Voting Classifier and Naïve bays were employed to forecast the occurrence of diabetic in patients. The findings of the study indicate that these machine learning algorithms achieved significant accuracy rates in diabetes prediction. Among the algorithms utilized, the Random Forest algorithm achieved the best accuracy rate of 86.5The study also discovered that a range of parameters, such as hypertension, age, body weight, and levels of glucose, were valid markers of diabetes. For individuals who have a greater chance of acquiring diabetes, these factors can help medical experts act early and provide unique treatment strategies.