Scholarlink

<p>Heart disease continues to be a prominent reason for death throughout the world, which highlights the importance of effective predictive models for better early diagnosis. Due to their limited scalability and inability to capture complex relationships, existing machine learning methods often need more accuracy when applied to larger datasets. SENNERF, a unique ensemble framework, integrates DenseNet topologies and an upgraded random forest model to improve heart disease prediction. This research proposes a scalable, accurate, and generalizable model for huge datasets. We perform feature extraction and enhanced random forest prediction after using DenseNet128, DenseNet256, and CompactNet64 architectures. SENNERF performed well on a large heart illness dataset, with 99.97% accuracy, 99% precision and recall, and 99.71% F1-score. This outperforms previous research that struggled with smaller datasets and had lower accuracy. SENNERF’s capacity to handle large datasets with high accuracy and cheap processing costs makes it a unique heart disease solution. Additionally, the proposed model was evaluated with the benchmark heart disease dataset. The proposed SENNERF model attained 100% accuracy for classifying the highly imbalanced small dataset. Click here to view <Emphasis Type="Underline">Source Code</Emphasis> of the study.</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

SENNERF: Stacking Ensemble with Neural Networks and Enhanced Random Forest for improved heart disease prediction

  • 期刊论文

  • 发表日期：

    2025年9月12日

查看全文
  • Muni Balaji Thumu,
  • N. Balajiraja,
  • Muhammed Yousoof

摘要

Heart disease continues to be a prominent reason for death throughout the world, which highlights the importance of effective predictive models for better early diagnosis. Due to their limited scalability and inability to capture complex relationships, existing machine learning methods often need more accuracy when applied to larger datasets. SENNERF, a unique ensemble framework, integrates DenseNet topologies and an upgraded random forest model to improve heart disease prediction. This research proposes a scalable, accurate, and generalizable model for huge datasets. We perform feature extraction and enhanced random forest prediction after using DenseNet128, DenseNet256, and CompactNet64 architectures. SENNERF performed well on a large heart illness dataset, with 99.97% accuracy, 99% precision and recall, and 99.71% F1-score. This outperforms previous research that struggled with smaller datasets and had lower accuracy. SENNERF’s capacity to handle large datasets with high accuracy and cheap processing costs makes it a unique heart disease solution. Additionally, the proposed model was evaluated with the benchmark heart disease dataset. The proposed SENNERF model attained 100% accuracy for classifying the highly imbalanced small dataset. Click here to view Source Code of the study.