Scholarlink

<p>The Internet of Medical Things (IoMT) is transforming modern healthcare by enabling real-time monitoring, intelligent diagnostics and efficient patient care. However, the interconnected nature of IoMT systems makes them highly vulnerable to cyber threats, highlighting the urgent need for robust intrusion detection systems (IDS). This research presents a detailed implementations and impact analysis of eight popular optimizers such as Adam, Adadelta, SGD, Adagrad, Adamax, AdamW, Nadam and RMSprop which are systematically applied to train and fine-tune Deep Neural Network (DNN) models using a MSCAD dataset reflective of IoMT network conditions. Using the MSCAD dataset, which closely reflects real-world IoMT network traffic, this research systematically evaluated the impact of each optimizer on key performance metrics, including precision, recall, F1-score, F2-score and Fβ-score. Among the tested configurations, the DNN model optimized with Adam demonstrated superior performance, achieving a precision of 0.9775, along with consistently high recall and F1-score. The experimental results underscore the significance of optimizer selection in training deep learning (DL)-based IDS for medical networks. The proposed DNN-Adam framework not only improves intrusion detection performance but also enhances the reliability and responsiveness of IoMT systems, thereby strengthening the protection of sensitive healthcare data. This research introduces valuable insights for designing effective cybersecurity solutions in resource-constrained, real-time medical environments, where both accuracy and efficiency are critical.</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

Optimizer Influenced Deep Neural Network for IoT Healthcare Security

  • 期刊论文

  • 发表日期：

    2026年1月19日

查看全文
  • Pooja Puspita Priyadarshani,
  • Janmenjoy Nayak,
  • Pandit Byomakesha Dash

摘要

The Internet of Medical Things (IoMT) is transforming modern healthcare by enabling real-time monitoring, intelligent diagnostics and efficient patient care. However, the interconnected nature of IoMT systems makes them highly vulnerable to cyber threats, highlighting the urgent need for robust intrusion detection systems (IDS). This research presents a detailed implementations and impact analysis of eight popular optimizers such as Adam, Adadelta, SGD, Adagrad, Adamax, AdamW, Nadam and RMSprop which are systematically applied to train and fine-tune Deep Neural Network (DNN) models using a MSCAD dataset reflective of IoMT network conditions. Using the MSCAD dataset, which closely reflects real-world IoMT network traffic, this research systematically evaluated the impact of each optimizer on key performance metrics, including precision, recall, F1-score, F2-score and Fβ-score. Among the tested configurations, the DNN model optimized with Adam demonstrated superior performance, achieving a precision of 0.9775, along with consistently high recall and F1-score. The experimental results underscore the significance of optimizer selection in training deep learning (DL)-based IDS for medical networks. The proposed DNN-Adam framework not only improves intrusion detection performance but also enhances the reliability and responsiveness of IoMT systems, thereby strengthening the protection of sensitive healthcare data. This research introduces valuable insights for designing effective cybersecurity solutions in resource-constrained, real-time medical environments, where both accuracy and efficiency are critical.