Explainable Machine Learning Framework for Cardiovascular Disease Diagnosis and Prognosis
Researchers have developed a unified machine learning framework designed to enhance the diagnosis and prognosis of cardiovascular diseases, addressing critical gaps in healthcare infrastructure. The study utilizes the Heart Disease dataset, comprising 1,035 instances, and employs SMOTE to generate 100,000 synthetic samples to mitigate data imbalance. The framework integrates classification techniques for disease detection and regression methods for risk forecasting. Evaluation results indicate that the Random Forest algorithm achieved superior performance in classification, with accuracy rates of 0.972 on real data and 0.976 on synthetic data. For prediction modeling, linear regression demonstrated the highest efficacy, yielding R2 values of 0.984 and 0.992 for real and synthetic samples, respectively, alongside minimal measurement errors. Additionally, Explainable AI methods were incorporated to improve the interpretability of model outcomes. This research highlights the transformative potential of machine learning in clinical settings, offering precise tools for timely interventions and improved management of heart disease risks globally.
Wire timeline
Explainable Machine Learning Framework for Cardiovascular Disease Diagnosis and Prognosis
Researchers have developed a unified machine learning framework designed to enhance the diagnosis and prognosis of cardiovascular diseases, addressing critical gaps in healthcare infrastructure. The study utilizes the Heart Disease dataset, comprising 1,035 instances, and employs SMOTE to generate 100,000 synthetic samples to mitigate data imbalance. The framework integrates classification techniques for disease detection and regression methods for risk forecasting. Evaluation results indicate that the Random Forest algorithm achieved superior performance in classification, with accuracy rates of 0.972 on real data and 0.976 on synthetic data. For prediction modeling, linear regression demonstrated the highest efficacy, yielding R2 values of 0.984 and 0.992 for real and synthetic samples, respectively, alongside minimal measurement errors. Additionally, Explainable AI methods were incorporated to improve the interpretability of model outcomes. This research highlights the transformative potential of machine learning in clinical settings, offering precise tools for timely interventions and improved management of heart disease risks globally.
cs.AI updates on arXiv.org