Abstract

Diabetes prediction plays a pivotal role in healthcare, offering numerous benefits for individuals and healthcare providers alike. In this study, we evaluate several machine learning algorithms for diabetes prediction, including Random Forest, Support Vector Machine (SVM), Gaussian Naive Bayes, Decision Tree, XGBoost, kNearest Neighbors, and Logistic Regression. Leveraging Python libraries such as pandas, scikitlearn, seaborn, and matplotlib, we preprocess the data, encode categorical variables, and standardize features to facilitate modeling. Through comprehensive experimentation and evaluation using train-test splitting, we find that XGBoost yields the highest accuracy of 97.17% among all algorithms. The ability to accurately predict diabetes offers numerous benefits, including early identification of individuals at risk, enabling timely intervention and lifestyle modifications to prevent or delay the onset of diabetes-related complications. Additionally, accurate prediction facilitates personalized healthcare interventions, optimizing treatment plans and resources allocation. This study underscores the potential of machine learning-based diabetes prediction in improving patient Outcomes and healthcare management.