Novel Pelican Optimization Algorithm (POA) With Stacked Sparse Autoencoder (SSAE) Based IDS for Network Security

ABSTRACT Security is a crucial factor for information systems and other vital infrastructures. Ensuring robust security measures is imperative due to the substantial volume of network traffic. On the other hand, many network components are susceptible to cyber threats and attacks due to their inhere...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Transactions on emerging telecommunications technologies Ročník 36; číslo 5
Hlavní autori: Kanimozhi, R., Neela Madheswari, A.
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Chichester, UK John Wiley & Sons, Ltd 01.05.2025
Predmet:
data privacy
deep learning
network security
optimization
pelican optimization algorithm
stacked sparse autoencoder
ISSN:2161-3915, 2161-3915
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:ABSTRACT Security is a crucial factor for information systems and other vital infrastructures. Ensuring robust security measures is imperative due to the substantial volume of network traffic. On the other hand, many network components are susceptible to cyber threats and attacks due to their inherent properties. The increasing use of networks paves the way for widespread security vulnerabilities. In this context, the implementation of intrusion detection systems (IDS) plays a key role in safeguarding information systems and their network architectures. This research introduces an optimized deep learning model aimed at improving network security by accurately detecting intrusions. The proposed IDS, also termed as the POA‐SSAE IDS model (pelican optimization model‐stacked sparse autoencoder), integrates a POA for optimal feature selection and an SSAE for feature classification. The effectiveness of this IDS was tested using benchmark datasets, namely CICIDS2018 and KDDCUP'99. The results exhibited the proposed model's superior performance, achieving an accuracy of 97.45% on the CICIDS2018 dataset and 98.7% accuracy on the KDDCUP'99 dataset. A novel intrusion detection framework that integrates the Pelican Optimization Algorithm (POA) with a Stacked Sparse Autoencoder (SSAE) for enhanced network security. By preprocessing raw traffic data from CICIDS2018 and KDDCUP'99 datasets through null removal, encoding, and normalization. POA is then applied to select the most relevant features, improving efficiency and reducing redundancy. These features are classified using SSAE, optimized with RMSprop to ensure accurate pattern recognition. The model delivers high detection accuracy of 97.45% on CICIDS2018 and 98.7% on KDDCUP'99 proving its effectiveness in detecting both known and emerging network threats.
ISSN:2161-3915
2161-3915
DOI:10.1002/ett.70113