Scholarlink

<p>To enhance IoT security, this paper presents a novel approach using deep learning model. This study introduces an advanced intrusion detection method that employs a Stacked Auto Encoder Network for efficient feature extraction, thereby reducing the complexity of the detection model. The extracted features are subsequently fed into anenhanced Capsule Networks (ECapsNets) model for attack detection, with weight parameters optimized using the Adaptive Osprey Optimization Algorithm (AO<sup>2</sup>). The method is evaluated using two datasets: UNSW-NB15, andNSL-KDD 99. Performance metrics like as detection rate, accuracy, precision, recall, and F-measure are utilized to assess the approach’s effectiveness. The proposed method achieved an accuracy of 99.6%, precision of 99.4%, recall of 98.4%, and an F-measure of 98.9% on the NSL-KDD 99 dataset. On the UNSW-NB15 dataset, it achieved an accuracy of 99.5%, precision of 99.1%, recall of 99%, and an F-measure of 98.9%. Experimental results demonstrate that the proposed method significantly improves detection accuracy and robustness against various security threats, thereby enhancing the overall security of IoT systems.</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

An optimal attack detection model based on optimized capsule networks with stacked auto encoder in IoT

  • 期刊论文

  • 发表日期：

    2025年6月20日

查看全文
  • T. Silambarasan,
  • Anitha Chikkanayakanahalli Lokesh Kumar,
  • N. V. Sanjay Kumar,
  • B. V. N. V. Krishna Suresh

摘要

To enhance IoT security, this paper presents a novel approach using deep learning model. This study introduces an advanced intrusion detection method that employs a Stacked Auto Encoder Network for efficient feature extraction, thereby reducing the complexity of the detection model. The extracted features are subsequently fed into anenhanced Capsule Networks (ECapsNets) model for attack detection, with weight parameters optimized using the Adaptive Osprey Optimization Algorithm (AO2). The method is evaluated using two datasets: UNSW-NB15, andNSL-KDD 99. Performance metrics like as detection rate, accuracy, precision, recall, and F-measure are utilized to assess the approach’s effectiveness. The proposed method achieved an accuracy of 99.6%, precision of 99.4%, recall of 98.4%, and an F-measure of 98.9% on the NSL-KDD 99 dataset. On the UNSW-NB15 dataset, it achieved an accuracy of 99.5%, precision of 99.1%, recall of 99%, and an F-measure of 98.9%. Experimental results demonstrate that the proposed method significantly improves detection accuracy and robustness against various security threats, thereby enhancing the overall security of IoT systems.