Deep Learning-Based Hybrid Algorithm for Detecting Cyber-Attacks
Ghazia Qaiser, Sivachandran Chandrasekaran, Rifai Chai, Jinchuan Zheng · 2024
The Cybersecurity of the Industrial Internet of Services (IIoS) requires substantial consideration due to its extensive nature of interconnected devices and data communication, which increasingly relies on wireless networks. The increased dependence of wireless networks and interconnected devices on the Internet increases the value of exploiting the IloS. The traditional methods used to detect cyber threats can no longer rely on a single aspect of information; instead, they must be able to adapt to new threats by learning from various sources. Deep learning has gained popularity in recent years for its ability to learn complex inferences from significant data sources. This research aims to propose a deep learning-based hybrid algorithm that combines Multi-Layer Perceptron (MLP) and Bidirectional Long Short-Term Memory (BiLSTM) networks. MLP and BiLSTM networks enhance the model's capability to detect complex cyber-attack patterns. Experimental results reveal better adequacy in terms of accuracy and recall than traditional methods. The findings suggested that the proposed hybrid model can be a valuable approach for detecting sophisticated attack vectors.