Scholarlink

<p>Recently, there has been a rising count of IoT devices linked to networks, and owing to the progression in technologies, cyber-attacks and security threats, like botnets, are evolving and emerging speedily with high risk. Such attacks interrupt IoT conversion by disturbing services and networks of IoT devices. A botnet attack is any attack leveraging a botnet, i.e. group of devices and bots associated together to carry out a similar task for scaling and distribution. Numerous current studies have adopted DL and ML schemes for classifying and detecting botnet attacks in IoT networks. In this way, a Botnet Attack Detection model in IoT (BAD-IoT) is introduced with the available dataset. In the first stage, pre-processing of data is done with improved tanh data normalization. In the subsequent stage, improved entropy, mutual information (MI) and statistical features are extracted. Subsequently, an ensemble model that includes models like Bi-GRU, LSTM and DMN models is deployed for the detection process, where the weights of Bi-GRU and LSTM are optimally elected via Butterfly Exploited Archimedes algorithm (BE-ARCH) during the training process. Finally, the comparison analysis is performed to evaluate the superiority of the proposed BAD-IoT model over the conventional models.</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

IoT botnet attack detection using ensemble classifiers with optimal training

  • 期刊论文

  • 发表日期：

    2026年4月20日

查看全文
  • Shankar Vuppu,
  • V. Chandra Shekhar Rao,
  • M. Sujatha,
  • Shaik Munawar,
  • S. Nagaraju,
  • N. Gayatri

摘要

Recently, there has been a rising count of IoT devices linked to networks, and owing to the progression in technologies, cyber-attacks and security threats, like botnets, are evolving and emerging speedily with high risk. Such attacks interrupt IoT conversion by disturbing services and networks of IoT devices. A botnet attack is any attack leveraging a botnet, i.e. group of devices and bots associated together to carry out a similar task for scaling and distribution. Numerous current studies have adopted DL and ML schemes for classifying and detecting botnet attacks in IoT networks. In this way, a Botnet Attack Detection model in IoT (BAD-IoT) is introduced with the available dataset. In the first stage, pre-processing of data is done with improved tanh data normalization. In the subsequent stage, improved entropy, mutual information (MI) and statistical features are extracted. Subsequently, an ensemble model that includes models like Bi-GRU, LSTM and DMN models is deployed for the detection process, where the weights of Bi-GRU and LSTM are optimally elected via Butterfly Exploited Archimedes algorithm (BE-ARCH) during the training process. Finally, the comparison analysis is performed to evaluate the superiority of the proposed BAD-IoT model over the conventional models.