Scholarlink

<p>Explainable machine learning is paramount in enhancing the transparency and interpretability of machine learning (ML) models, especially within the medical domain. This study employs six different ML algorithms to develop a predictive model for heart disease by utilizing a specific dataset dedicated to this condition. To augment interpretability, we integrated two prominent explainable techniques: SHapley Additive exPlanations and Local Interpretable Model-agnostic Explanations. The primary aim was to furnish medical practitioners with a dependable tool for the early detection and prediction of heart disease, thereby enhancing treatment efficacy and improving patient outcomes. The performance of the deployed models was appraised using several metrics: accuracy, recall, precision, ROC-AUC, and F1-Score. Notably, the Logistic Regression model outperformed others, attaining an accuracy of 87%, recall of 91%, precision of 82%, ROC-AUC of 92%, and an F1-Score of 86%. These findings hold significant implications for medical practitioners, potentially improving heart disease detection and patient care management.</p>

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

Implementing a Heart Disease Prediction Model with Explainable Machine Learning Techniques

  • 期刊论文

  • 发表日期：

    2025年9月29日

查看全文
  • Krishna Mridha,
  • Madhu Shukla,
  • Biswaranjan Acharya,
  • Vassilis C. Gerogiannis,
  • Andreas Kanavos,
  • Masrur Ahsan Priyok,
  • Rayhan Hussain Razu

摘要

Explainable machine learning is paramount in enhancing the transparency and interpretability of machine learning (ML) models, especially within the medical domain. This study employs six different ML algorithms to develop a predictive model for heart disease by utilizing a specific dataset dedicated to this condition. To augment interpretability, we integrated two prominent explainable techniques: SHapley Additive exPlanations and Local Interpretable Model-agnostic Explanations. The primary aim was to furnish medical practitioners with a dependable tool for the early detection and prediction of heart disease, thereby enhancing treatment efficacy and improving patient outcomes. The performance of the deployed models was appraised using several metrics: accuracy, recall, precision, ROC-AUC, and F1-Score. Notably, the Logistic Regression model outperformed others, attaining an accuracy of 87%, recall of 91%, precision of 82%, ROC-AUC of 92%, and an F1-Score of 86%. These findings hold significant implications for medical practitioners, potentially improving heart disease detection and patient care management.