Scholarlink

<p>The rapid growth of the Internet of Things (IoT) devices has led to a surge in cyber threats, making it crucial to develop advanced Intrusion Detection Systems (IDS) to safeguard network security and reliability. This research presents an intelligent IDS framework designed to enhance threat detection in IoT environments. The core idea behind this work is that integrating deep learning techniques, such as convolutional neural networks (CNNs) for feature extraction, attention-based bidirectional long-short-term memory (BiLSTM) for analyzing time-based patterns, and artificial neural networks (ANNs) for classification can significantly enhance the accuracy and efficiency of IDS. The proposed approach follows a structured methodology, starting with the handling of class imbalance using a hybrid method that combines Adaptive Synthetic Sampling (ADASYN) and One-Sided Selection (OSS) techniques. Additionally, feature selection is employed using CNNs to extract the most relevant features while eliminating redundant information, thereby optimizing computational efficiency. Next, BiLSTM, enhanced with an attention mechanism, is employed to capture sequential dependencies in network traffic data, ensuring accurate detection of cyber threats. Finally, ANNs are utilized to classify network traffic as normal or malicious, allowing the system to adapt effectively to diverse cyberattack patterns. Extensive evaluations were conducted using benchmark IDS datasets, including NSL-KDD and UNSW-NB15, to assess the effectiveness of the framework. The experimental results demonstrate that the proposed model achieves a remarkable 99% accuracy on the NSL-KDD dataset and 94% accuracy on the UNSW-NB15 dataset, outperforming existing IDS approaches. Furthermore, the model effectively addresses critical challenges such as class imbalance and high false positive rates, making it a promising solution to strengthen IoT security.</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

Attention-enhanced BiLSTM-ANN framework with CNN-based feature selection for advanced threat detection

  • 期刊论文

  • 发表日期：

    2026年1月29日

查看全文
  • Mohammed Tayebi,
  • Said El Kafhali

摘要

The rapid growth of the Internet of Things (IoT) devices has led to a surge in cyber threats, making it crucial to develop advanced Intrusion Detection Systems (IDS) to safeguard network security and reliability. This research presents an intelligent IDS framework designed to enhance threat detection in IoT environments. The core idea behind this work is that integrating deep learning techniques, such as convolutional neural networks (CNNs) for feature extraction, attention-based bidirectional long-short-term memory (BiLSTM) for analyzing time-based patterns, and artificial neural networks (ANNs) for classification can significantly enhance the accuracy and efficiency of IDS. The proposed approach follows a structured methodology, starting with the handling of class imbalance using a hybrid method that combines Adaptive Synthetic Sampling (ADASYN) and One-Sided Selection (OSS) techniques. Additionally, feature selection is employed using CNNs to extract the most relevant features while eliminating redundant information, thereby optimizing computational efficiency. Next, BiLSTM, enhanced with an attention mechanism, is employed to capture sequential dependencies in network traffic data, ensuring accurate detection of cyber threats. Finally, ANNs are utilized to classify network traffic as normal or malicious, allowing the system to adapt effectively to diverse cyberattack patterns. Extensive evaluations were conducted using benchmark IDS datasets, including NSL-KDD and UNSW-NB15, to assess the effectiveness of the framework. The experimental results demonstrate that the proposed model achieves a remarkable 99% accuracy on the NSL-KDD dataset and 94% accuracy on the UNSW-NB15 dataset, outperforming existing IDS approaches. Furthermore, the model effectively addresses critical challenges such as class imbalance and high false positive rates, making it a promising solution to strengthen IoT security.