Scholarlink

Heart diseases are among the highest causes of World’s mortality globally. They diseases are classified into different varieties like vascular, ischemic, and hypertensive heart disease. Many medical attributes have been documented among the patients within Electronic Health Records (EHR) that can aid doctors to monitor and diagnose the disease of the heart. In this paper, the generative AI-powered ensemble learning solution to diagnose the risk factors of the disease of the heart with the support of the Behavior Risk Factor Surveillance System (BRFSS) database has been introduced. The database encompasses the key healthcare indicators such as BMI, current smoking status, exercise levels, and comorbid conditions like diabetes. To overcome the problem of the skewed classes of the database, synthetic augmentation of the database is implemented with the support of a Generative Adversarial Network (GAN), followed by ensemble classification that uses the models of the ensemble of LightGBM, Random Forest, and Neural Networks. Performance metrics like accuracy, precision, recall, and AUC-ROC support the claim that the proposed solution outperforms the baselines with a 94.75% rate of accuracy.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Classification of Heart Disease Risk Factors Using an Ensemble Model with Generative AI Augmentation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ronit Mathur,
  • Shilpa Gupta,
  • Deepika Kumar,
  • Akhtar Jamil

摘要

Heart diseases are among the highest causes of World’s mortality globally. They diseases are classified into different varieties like vascular, ischemic, and hypertensive heart disease. Many medical attributes have been documented among the patients within Electronic Health Records (EHR) that can aid doctors to monitor and diagnose the disease of the heart. In this paper, the generative AI-powered ensemble learning solution to diagnose the risk factors of the disease of the heart with the support of the Behavior Risk Factor Surveillance System (BRFSS) database has been introduced. The database encompasses the key healthcare indicators such as BMI, current smoking status, exercise levels, and comorbid conditions like diabetes. To overcome the problem of the skewed classes of the database, synthetic augmentation of the database is implemented with the support of a Generative Adversarial Network (GAN), followed by ensemble classification that uses the models of the ensemble of LightGBM, Random Forest, and Neural Networks. Performance metrics like accuracy, precision, recall, and AUC-ROC support the claim that the proposed solution outperforms the baselines with a 94.75% rate of accuracy.