Abstract :

Coronary Heart Disease (CHD) Is A Major Cardiovascular Condition For Which Accurate And Timely Prediction Can Support Early Clinical Intervention And Improve Patient Management. Conventional Machine Learning Approaches May Experience Limitations Related To Class Imbalance, Parameter Selection, And Variations In Predictive Performance. This Study Presents An Ensembleenhanced Framework For CHD Prediction That Combines An Optimized LightGBM Classifier With Ensemble Learning Techniques. The Proposed Approach Utilizes CHD Data From The Framingham Heart Institute And Performs Data Preprocessing, Missingvalue Handling, Normalization, Feature Preparation, And Train-test Data Partitioning. LightGBM Is Optimized Through Hyperparameter Tuning Using The Optuna Framework, While Focal Loss Is Incorporated To Improve Learning Under Imbalanced Class Distributions. As An Extension, A Voting Classifier Integrating Random Forest And AdaBoost Is Introduced To Combine The Complementary Predictive Capabilities Of Multiple Classifiers And Improve Classification Robustness. The Developed System Achieves A Reported Accuracy Of 99% With The RF–AdaBoost Voting Classifier, Demonstrating Improved Predictive Performance Compared With Individual Classification Approaches. The Framework Is Further Integrated With A Flask-based User Interface And SQLite Database To Support User Registration, Authentication, Input Submission, And Prediction Testing. The Proposed Ensemble-based Framework Provides An Effective And Practical Approach For Automated CHD Prediction And Can Serve As A Supportive Tool For Early Cardiovascular Risk Assessment.

Published:

20-8-2026

Issue:

Vol. 26 No. 8 (2026)

Page Nos:

1023-1032

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite