Scholarlink

This paper presents a comprehensive framework for detecting anomalies in IoT-based healthcare systems, where integrating IoT technologies has improved patient care but also heightened cybersecurity vulnerabilities. The proposed framework leverages a Stacking Classifier, utilizing Random Forest and Gradient Boosting as base learners, while Logistic Regression serves as the meta-classifier for final predictions. The experimental results, evaluated on two publicly available healthcare datasets, demonstrate an average accuracy rate of 99.75%, confirming the framework’s efficacy in classifying malicious and benign data. This work underscores the importance of real-time anomaly detection in improving the security and reliability of IoT-based healthcare environments, providing a significant contribution toward robust solutions in this domain. Future work will be focused on applying deep learning techniques to enhance the capabilities of the proposed framework.

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

Enhancing Anomaly Detection in IoT Systems with an Ensemble of Machine Learning Algorithms

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ishika Agarwal,
  • Krishna Girish,
  • Prarthna Puhan,
  • Ramanathan Lakshmanan

摘要

This paper presents a comprehensive framework for detecting anomalies in IoT-based healthcare systems, where integrating IoT technologies has improved patient care but also heightened cybersecurity vulnerabilities. The proposed framework leverages a Stacking Classifier, utilizing Random Forest and Gradient Boosting as base learners, while Logistic Regression serves as the meta-classifier for final predictions. The experimental results, evaluated on two publicly available healthcare datasets, demonstrate an average accuracy rate of 99.75%, confirming the framework’s efficacy in classifying malicious and benign data. This work underscores the importance of real-time anomaly detection in improving the security and reliability of IoT-based healthcare environments, providing a significant contribution toward robust solutions in this domain. Future work will be focused on applying deep learning techniques to enhance the capabilities of the proposed framework.