Abstract
This paper presents the implementation and experimental evaluation of an intelligent machine learning-based diabetes prediction system using the Pima Indians Diabetes Dataset. The system is designed around an XGBoost classifier and includes data preprocessing, feature preparation, stratified train-test partitioning, model training, prediction, performance evaluation and feature-importance analysis. The dataset contains records and eight clinical predictors: Pregnancies, Glucose, BloodPressure, SkinThickness, Insulin, BMI, DiabetesPedigreeFunction and Age. Five machine learning models—Logistic Regression, Decision Tree, Random Forest, Support Vector Machine and XGBoost—were implemented for comparative evaluation. The models were evaluated using Accuracy, Precision, Recall, F1-Score, ROC-AUC and Confusion Matrix. XGBoost feature-importance analysis identifies the clinical variables contributing most strongly to prediction.