Prediction of Heart Disease Using XGBoost Algorithm

Heart disease remains a major global health concern, making early diagnosis crucial for reducingmortality rates. This study focuses on developing a predictive model using the XGBoost algorithm,optimized through Optuna hyperparameter tuning, and ANOVA-based feature selection for identifyingthe most important clinical indicators. The Framingham Heart Study dataset is utilized, incorporatingkey cardiovascular risk factors such as age, blood pressure, cholesterol, BMI, and glucose levels. Theapproach includes ANOVA-based feature selection to identify the most relevant predictors andXGBoost with Optuna hyperparameter tuning to enhance predictive accuracy.To improve model performance, data preprocessing techniques such as handling missing values,applying Standard Scaler for feature scaling, and addressing class imbalance using Synthetic MinorityOver-sampling Technique (SMOTE) were implemented. The optimized model demonstrates highaccuracy, making it a reliable tool for risk assessment. By leveraging advanced machine learningtechniques, this model can assist healthcare professionals in making informed decisions, ultimatelyaiding in early detection and preventive care for heart disease. The final XGBoost classifier achieved agood accuracy score, demonstrating its effectiveness in predicting heart disease risk. To make thismodel accessible for real-world use, an interactive web application was developed allowing users toinput clinical parameters and receive instant heart disease risk predictions.