Enhancing Iot Network Security Using Neural Network Approach for Intrusion Detection
Hye Jin Kim, Rhee Jung Soo · Asia-pacific Journal of Convergent Research Interchange · 2024
In the dynamic landscape of Internet of Things (IoT) networks, robust security is imperative to counter the escalating complexity and scale of cyber-attacks.This work introduces an innovative deep learning-based intrusion detection system, aiming to enhance the security of IoT-based networks.Acknowledging the limitations of traditional security measures, the study underscores the pivotal role of data preprocessing in ensuring its suitability for effective learning and accurate intrusion detection.The preprocessing module encompasses data cleaning, one-hot encoding, and normalization, generating well-prepared inputs for the subsequent deep-learning architecture.The proposed architecture features an efficient Artificial Neural Network (ANN) framework, integrating multilayered neural networks to autonomously learn and extract essential features.This empowers the system to identify unauthorized access attempts and potential malware attacks in an automated and efficient manner.The study examines the outcomes of the suggested research employing the UNSW-NB15 incursion dataset, comprising 82,332 data points.Thirty percent of the dataset was put aside for testing, while the remaining seventy percent was used for training.This partitioning allowed for a comprehensive evaluation of the proposed approach's performance on both preparation and analysis data.The suggested study significantly contributes to fortifying the safety and resilience of IoT networks against evolving threats and vulnerabilities.