PlumX Metrics
Embed PlumX Metrics

Detecting Network Intrusions with Resilient Approaches Based on Convolutional Neural Networks

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), ISSN: 1611-3349, Vol: 13798 LNCS, Page: 238-248
2023
  • 0
    Citations
  • 0
    Usage
  • 2
    Captures
  • 0
    Mentions
  • 0
    Social Media
Metric Options:   Counts1 Year3 Year

Metrics Details

Conference Paper Description

An anomaly-dependent network intrusion detection method is proposed by presenting a hybrid of Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) approaches where it detects various network attacks by training a CNN on the NSL-KDD dataset. The proposed approach is able to automatically discovers significant traits without human intervention and achieves higher accuracy in detecting attacks which can improve the performance of intrusion detection systems. The CNN-LSTM approach is applied to detect attacks such as DoS, Probe, U2R, and R2L, and identify normal traffic at the same time. The experimental result reached a low false-positive rate, a high accuracy rate, and a low time average. The results show that the proposed approach outperformed the current state-of-the-art by 4.69% in terms of accuracy and conventional CNN achieved a greater accuracy of roughly 3% compared to the current approach, which amounted to 98.95%, 97.81%, and 94.26% respectively. The consistency of the simulated attacks is unequal which leads to issues in the prediction, therefore, additional efforts to collect more consistent traffic on unusual attacks are required. The proposed approach is targeted to be applied as an effective tool for solving complex classification problems such as NIDS.

Provide Feedback

Have ideas for a new metric? Would you like to see something else here?Let us know