Scholarlink

The objective of this paper was to use machine learning (ML) and deep learning (DL) approaches to identify chronic heart failure (HF). We were able to replicate the method utilized in clinical practice since the models we built in this work are based on many combinations of feature categories, including clinical features, echocardiography data, and laboratory findings. The incremental value for each feature category was also looked at. There were 422 subjects used in all. A suggested machine learning method includes steps for feature selection, class imbalance management, and classification. When features from all categories were included, the results for the diagnosis of HF were quite satisfactory, with high accuracy (97.56%), sensitivity (100%), and specificity (96.55%). Even with the single feature type used, the results were still quite good.

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

Detection of Chronic Heart Failure Using Deep Learning and Machine Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • P. H. V. Sesha Talpa Sai,
  • V. V. Pratibha Bharathi,
  • Umeshbabu Nandanan,
  • Suraj Raju Chikode,
  • Mrutyunjaya Munnolimath,
  • Kini Bhanu Prakash,
  • G. S. Naveen Kumar,
  • Amiya Bhaumik

摘要

The objective of this paper was to use machine learning (ML) and deep learning (DL) approaches to identify chronic heart failure (HF). We were able to replicate the method utilized in clinical practice since the models we built in this work are based on many combinations of feature categories, including clinical features, echocardiography data, and laboratory findings. The incremental value for each feature category was also looked at. There were 422 subjects used in all. A suggested machine learning method includes steps for feature selection, class imbalance management, and classification. When features from all categories were included, the results for the diagnosis of HF were quite satisfactory, with high accuracy (97.56%), sensitivity (100%), and specificity (96.55%). Even with the single feature type used, the results were still quite good.