CNN-LSTM Hybrid Architecture for Accurate Network Intrusion Detection for Cybersecurity

Fuente: Zenodo
Saved in:
Bibliographic Details
Main Authors: Dinesh Rajendran, Vaibhav Maniar, Vetrivelan Tamilmani, Venkata Deepak Namburi, Aniruddha Arjun Singh Singh, Rami Reddy Kothamaram
Format: Recurso digital
Published: Zenodo 2023
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866901565558751232
author Dinesh Rajendran
Vaibhav Maniar
Vetrivelan Tamilmani
Venkata Deepak Namburi
Aniruddha Arjun Singh Singh
Rami Reddy Kothamaram
author_facet Dinesh Rajendran
Vaibhav Maniar
Vetrivelan Tamilmani
Venkata Deepak Namburi
Aniruddha Arjun Singh Singh
Rami Reddy Kothamaram
contents <p><span>Cyber threat network security is of importance because the technology is progressively turning out to be important in the era of heightened digitalization. Improving cyber intrusion detection systems is crucial because cyberattacks on critical infrastructure are becoming more targeted and sophisticated. Using the NSL-KDD dataset for IDS is best accomplished with a CNN-LSTM model, according to the research. The workflow contains expensive preprocessing such as label encoding, feature selection, and z-score normalization, and 70 30 train-test splits. The spatial features of the traffic data are learnt with CNN and the sequential dependencies are performed by LSTM, which allows the robust classification of normal and attack traffic. The suggested model outperforms standard classifiers like SVM and LR experimentally, with results of 94.32% accuracy, 99.32% precision, 97.88% recall, and 98.60% F1-score.<span>  </span>Based on these findings, CNN-LSTM hybrid is an intriguing method for enhancing detection accuracy while decreasing false positives.<span>  </span>Deep learning-based intruder detection systems could help improve cybersecurity, but further research is needed to evaluate their performance on new data, address class imbalance, and enable real-time operation, as the study suggests.</span></p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_17516985
institution Zenodo
language
publishDate 2023
publisher Zenodo
record_format zenodo
spellingShingle CNN-LSTM Hybrid Architecture for Accurate Network Intrusion Detection for Cybersecurity
Dinesh Rajendran
Vaibhav Maniar
Vetrivelan Tamilmani
Venkata Deepak Namburi
Aniruddha Arjun Singh Singh
Rami Reddy Kothamaram
<p><span>Cyber threat network security is of importance because the technology is progressively turning out to be important in the era of heightened digitalization. Improving cyber intrusion detection systems is crucial because cyberattacks on critical infrastructure are becoming more targeted and sophisticated. Using the NSL-KDD dataset for IDS is best accomplished with a CNN-LSTM model, according to the research. The workflow contains expensive preprocessing such as label encoding, feature selection, and z-score normalization, and 70 30 train-test splits. The spatial features of the traffic data are learnt with CNN and the sequential dependencies are performed by LSTM, which allows the robust classification of normal and attack traffic. The suggested model outperforms standard classifiers like SVM and LR experimentally, with results of 94.32% accuracy, 99.32% precision, 97.88% recall, and 98.60% F1-score.<span>  </span>Based on these findings, CNN-LSTM hybrid is an intriguing method for enhancing detection accuracy while decreasing false positives.<span>  </span>Deep learning-based intruder detection systems could help improve cybersecurity, but further research is needed to evaluate their performance on new data, address class imbalance, and enable real-time operation, as the study suggests.</span></p>
title CNN-LSTM Hybrid Architecture for Accurate Network Intrusion Detection for Cybersecurity
url https://doi.org/10.5281/zenodo.17516985