The proposed hybrid deep learning intrusion prediction IoT (HDLIP-IoT) framework

Area of Science:

• Cybersecurity
• Network Security
• Artificial Intelligence

Background:

• The Internet of Things (IoT) is rapidly expanding, increasing its vulnerability to cyber threats.
• Distributed Denial of Service (DDoS) attacks are a primary concern for IoT security.
• Existing DDoS detection methods struggle with accuracy due to evolving attack patterns.

Purpose of the Study:

• To develop an accurate and efficient method for detecting DDoS attacks in IoT environments.
• To leverage advanced machine learning techniques for enhanced network security.
• To address the limitations of current DDoS detection systems.

Main Methods:

• Utilized the Modified Whale Optimization Algorithm (MWOA) for optimal feature extraction from IP packets.
• Implemented a Hybrid Long Short Term Memory (LSTM) neural network for attack detection.
• Optimized LSTM weights using MWOA to minimize prediction errors.

Main Results:

• The proposed MWOA-LSTM model demonstrated superior performance in detecting DDoS attacks.
• Achieved higher precision, recall, and accuracy compared to standard Support Vector Machines (SVM) and Genetic Algorithm (GA).
• Effectively identified DDoS attacks by analyzing network traffic patterns.

Conclusions:

• The MWOA-LSTM framework offers a robust solution for real-time DDoS attack detection in IoT.
• This hybrid approach significantly improves the accuracy and efficiency of cybersecurity measures.
• The study highlights the potential of MWOA and LSTM in advancing network threat identification.