Scholarlink

Heart disease and the risk of heart attack have increased recently. Prognosing heart illness is difficult for doctors. Computer-assisted diagnosis research is persistent and growing. Computer-assisted diagnostics uses AI extensively. AI could provide elegant and automated methods for the study of high-dimensional, multimodal biological data used by biomedical experts to diagnose and understand disease. The study uses an Artificial Neural Network (ANN) to forecast a patient's likelihood of having coronary heart disease. A few of the performance indicators have been used to evaluate the prediction model. As AI grows more significant in healthcare, concerns about its explainability, transparency, and model bias are growing. This makes Explainable Artificial Intelligence (XAI) relevant. XAI promotes the adoption of AI in healthcare by increasing the confidence of medical practitioners and AI researchers in an AI system. This work describes the application of Explainable AI (SHapley Additive exPlanations—SHAP) to find informative features and their interpretations from the ANN model used for predicting coronary heart disease. The ANN model has given extremely good performance with an accuracy of 0.99, AUC of 1.0 and recall of 1.0 which means that all the patients having heart disease were predicted positive. The model when interpreted using SHAP, clearly shows the significance of each feature and their influence in predicting the presence of heart disease in a person. Also, while interpreting the prediction for an individual, the plots show based on which of the features the person has been predicted as a positive or negative case.

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

Identification of Key Biomarkers Impacting the Prediction of Heart Disease Using Artificial Neural Networks and Model Interpretability with SHAP

  • 章节

  • 发表日期：

    2026年

查看全文
  • G. Ramanathan,
  • S. N. Jagadeesha

摘要

Heart disease and the risk of heart attack have increased recently. Prognosing heart illness is difficult for doctors. Computer-assisted diagnosis research is persistent and growing. Computer-assisted diagnostics uses AI extensively. AI could provide elegant and automated methods for the study of high-dimensional, multimodal biological data used by biomedical experts to diagnose and understand disease. The study uses an Artificial Neural Network (ANN) to forecast a patient's likelihood of having coronary heart disease. A few of the performance indicators have been used to evaluate the prediction model. As AI grows more significant in healthcare, concerns about its explainability, transparency, and model bias are growing. This makes Explainable Artificial Intelligence (XAI) relevant. XAI promotes the adoption of AI in healthcare by increasing the confidence of medical practitioners and AI researchers in an AI system. This work describes the application of Explainable AI (SHapley Additive exPlanations—SHAP) to find informative features and their interpretations from the ANN model used for predicting coronary heart disease. The ANN model has given extremely good performance with an accuracy of 0.99, AUC of 1.0 and recall of 1.0 which means that all the patients having heart disease were predicted positive. The model when interpreted using SHAP, clearly shows the significance of each feature and their influence in predicting the presence of heart disease in a person. Also, while interpreting the prediction for an individual, the plots show based on which of the features the person has been predicted as a positive or negative case.