An intelligent deep representation learning with enhanced feature selection approach for cyberattack detection in internet of things enabled cloud environment

Hayam Alamro ¹, Sami Saad Albouq ², Jahangir Khan ³

⁰Department of Information Systems, College of Computer and Information Sciences, Princess Nourah bint Abdulrahman University, P.O. Box 84428, 11671, Riyadh, Saudi Arabia.

Scientific Reports +

|

September 30, 2025

View abstract on PubMed

Summary

This study introduces an intelligent hybrid deep learning method to enhance cybersecurity in cloud and Internet of Things (IoT) networks. The approach effectively detects cyberattacks with high accuracy, improving network security.

Area Of Science

Cybersecurity
Artificial Intelligence
Cloud Computing and IoT

Background

Cloud computing (CC) and Internet of Things (IoT) systems are increasingly vital but face sophisticated cyberattacks.
Traditional intrusion detection systems (IDS) struggle with advanced malware and evasion techniques.
Artificial intelligence (AI), particularly deep learning (DL), offers advanced solutions for detecting complex cyber threats.

Purpose Of The Study

To propose an Intelligent Hybrid Deep Learning Method for Cyber Attack Detection Using an Enhanced Feature Selection Technique (IHDLM-CADEFST) for IoT-enabled cloud networks.
To improve the cybersecurity posture of IoT systems by identifying critical threats and developing robust detection strategies.
To enhance the accuracy and efficiency of cyberattack detection in complex network environments.

Main Methods

Data pre-processing using the standard scaler method.
Feature selection (FS) employing recursive feature elimination with information gain (RFE-IG) to identify relevant features and prevent overfitting.
Implementation of a hybrid Convolutional Neural Network and Long Short-Term Memory (CNN-LSTM) model with the RMSprop optimizer for attack classification.

Main Results

The IHDLM-CADEFST approach demonstrated superior performance on the ToN-IoT and Edge-IIoT datasets.
Achieved high accuracy rates of 99.45% on ToN-IoT and 99.19% on Edge-IIoT.
Outperformed recent state-of-the-art models in cyberattack detection.

Conclusions

The proposed IHDLM-CADEFST method significantly enhances cybersecurity in IoT-enabled cloud networks.
The hybrid DL model combined with effective feature selection provides a powerful tool for detecting sophisticated cyber threats.
This research contributes to strengthening the security and privacy of digital assets in interconnected systems.

Keywords:

Cloud computing Cyber attack detection Cybersecurity Feature selection Information gain Internet of things Recursive feature elimination