Machine Learning-Based Stacking Ensemble Model for Prediction of Heart Disease with Explainable AI and K-Fold Cross-Validation: A Symmetric Approach

Symmetry, Journal Year: 2025, Volume and Issue: 17(2), P. 185 - 185

Published: Jan. 25, 2025

One of the most complex and prevalent diseases is heart disease (HD). It among main causes death around globe. With changes in lifestyles environment, its prevalence rising rapidly. The prediction early stages crucial, as delays diagnosis can cause serious complications even death. Machine learning (ML) be effective this regard. Many researchers have used different techniques for efficient detection to overcome drawbacks existing models. Several ensemble models also been applied. We proposed a stacking model named NCDG, which uses Naive Bayes, Categorical Boosting, Decision Tree base learners, with Gradient Boosting serving meta-learner classifier. performed preprocessing using factorization method convert string columns into integers. employ Synthetic Minority Oversampling TEchnique (SMOTE) BorderLineSMOTE balancing address issue data class imbalance. Additionally, we implemented hard soft voting classifier compared results model. For Artificial Intelligence-based eXplainability our NCDG model, use SHapley Additive exPlanations (SHAP) technique. outcomes show that suggested performs better than benchmark techniques. experimental achieved highest accuracy, F1-Score, precision recall 0.91, 0.91 respectively, an execution time 653 s. Moreover, utilized K-Fold Cross-Validation validate predicted results. worth mentioning their validation strongly coincide each other proves approach symmetric.

Language: Английский

Enhancing Trauma Care: A Machine Learning Approach with XGBoost for Predicting Urgent Hemorrhage Interventions Using NTDB Data

Bioengineering, Journal Year: 2024, Volume and Issue: 11(8), P. 768 - 768

Published: July 30, 2024

Trauma is a leading cause of death worldwide, with many incidents resulting in hemorrhage before the patient reaches hospital. Despite advances trauma care, majority deaths occur within first three hours hospital admission, offering very limited window for effective intervention. Unfortunately, significant increase mortality from hemorrhagic primarily due to delays control. Therefore, we propose machine learning model predict need urgent

Language: Английский