|
AUTHOR(S): Amjad Jumaah Frhan
|
TITLE SmartShield: A CGAN-Boosted Model for Detecting IoT Cyber Threats |
 PDF |
ABSTRACT As the Internet of Things (IoT) has evolved very quickly, serious security concerns have cropped up, with ensuring linked devices emerging vital. This study presents a data-driven system that combines synthetic data generation and machine learning (ML) to accurately detect cyberattacks on IoT devices. It uses CGAN technology to generate synthetic attack data and then uses LightGBM to identify attack patterns. The approach includes cleaning of IoT data and a LightGBM feature choice strategy. It learns to distinguish between different types of attacks, such as DoS attacks, ARP poisoning attacks, and data theft attacks. Additionally, it makes use of a gradient boosting architecture, which strikes a useful balance between computing cost and detection accuracy. Moreover, the suggested model outperformed previous intrusion detection models with an accuracy of 87% in detecting attacks on the RT-IoT2022 dataset supplemented with GAN-generated data. These outcomes say that the tree-guided gradient strengthening procedure may greatly decrease the expense of processing and has many possibilities for deployment in IoT contexts. Studies suggest the intrusion detection efficiency of assaults provide use with a useful tool for boosting businesses' safety ratings in an increasingly linked and prone electronic zoon. |
KEYWORDS Intrusion detection systems (IDS), Cyber-Security, Machine learning, IoT, LightGBM Model |
|
Cite this paper Amjad Jumaah Frhan. (2025) SmartShield: A CGAN-Boosted Model for Detecting IoT Cyber Threats. International Journal of Computers, 10, 294-302 |
|
