Scholarlink

<p>With the continuous rise of different network attacks, ensuring network security has become increasingly crucial in today’s digital world. Intrusion detection is a significant approach, widely employed for ensuring network security. However, traditional detection methods are limited by processing the imbalanced data, leading to challenges such as poor decision-making and reduced performance in accurately classifying network attacks. To address these challenges, a novel approach named the Hybrid Partition weighted resampler-based Difficult Set Sampling Technique + Spinal Siamese Convolutional Neural Network (PWRe_DSSTE + 2SCNNet) is introduced for intrusion detection in Imbalanced Network Traffic. In the proposed approach, the imbalanced input data is preprocessed with Z-score normalization. Subsequently, data imbalance is handled by the PWRe_DSSTE algorithm. After that, the 2SCNNet is utilized for intrusion detection in the balanced data. The 2SCNNet is developed with the integration of Siamese Convolutional Neural Network (SCNN) and SpinalNet, where the layers are modified. The evaluation results shows that the PWRe_DSSTE + 2SCNNet achieved an accuracy of 98.875%, precision of 95.047%, recall of 95.043%, an F1-score of 95.045% and kappa coefficient of 95.580%.</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

2SCNNet: spinal siamese convolutional neural network for intrusion detection in imbalanced network traffic

  • 期刊论文

  • 发表日期：

    2025年8月11日

查看全文
  • S. Suresh,
  • P. Babu

摘要

With the continuous rise of different network attacks, ensuring network security has become increasingly crucial in today’s digital world. Intrusion detection is a significant approach, widely employed for ensuring network security. However, traditional detection methods are limited by processing the imbalanced data, leading to challenges such as poor decision-making and reduced performance in accurately classifying network attacks. To address these challenges, a novel approach named the Hybrid Partition weighted resampler-based Difficult Set Sampling Technique + Spinal Siamese Convolutional Neural Network (PWRe_DSSTE + 2SCNNet) is introduced for intrusion detection in Imbalanced Network Traffic. In the proposed approach, the imbalanced input data is preprocessed with Z-score normalization. Subsequently, data imbalance is handled by the PWRe_DSSTE algorithm. After that, the 2SCNNet is utilized for intrusion detection in the balanced data. The 2SCNNet is developed with the integration of Siamese Convolutional Neural Network (SCNN) and SpinalNet, where the layers are modified. The evaluation results shows that the PWRe_DSSTE + 2SCNNet achieved an accuracy of 98.875%, precision of 95.047%, recall of 95.043%, an F1-score of 95.045% and kappa coefficient of 95.580%.