Scholarlink

With the increase in cyber threats, accurate Intrusion Detection Systems (IDS) for the network are essential for protection. This paper introduces a hybrid model that combines Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) layers to detect intrusions, using the CICIDS 2019 dataset, which includes attacks such as Botnet, DDoS, and DoS. Preprocessing involves the normalization and encoding of the time-series input for the deep learning approach. In the proposed model architecture, CNN is used for spatial feature extraction while LSTM is used for temporal relationships, with the test accuracy of the model being somewhere around 99.93%. The use of the model in the context of SIoT shows that it can help improve IDS performance in the IoT environment with high-quality spatial–temporal analysis of patterns.

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

Efficient Intrusion Detection in Social IoT: A Hybrid CNN-LSTM Model

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • S. Divya,
  • R. Tanuja

摘要

With the increase in cyber threats, accurate Intrusion Detection Systems (IDS) for the network are essential for protection. This paper introduces a hybrid model that combines Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) layers to detect intrusions, using the CICIDS 2019 dataset, which includes attacks such as Botnet, DDoS, and DoS. Preprocessing involves the normalization and encoding of the time-series input for the deep learning approach. In the proposed model architecture, CNN is used for spatial feature extraction while LSTM is used for temporal relationships, with the test accuracy of the model being somewhere around 99.93%. The use of the model in the context of SIoT shows that it can help improve IDS performance in the IoT environment with high-quality spatial–temporal analysis of patterns.