Comparison Analysis of Intrusion Detection using Deep Learning in IoT Networks

This study focuses on network security for Internet of Things (IoT) devices, which are particularly vulnerable to attack due to limited security measures. To protect IoT networks from attackers, the research implemented intrusion detection on the networks. The study compared several Deep Learning algorithms, including DNN, CNN, LSTM, and AE, to identify the most effective algorithm for solving network security problems using intrusion detection. The research used the UNSW-NB15 dataset for testing and employed binary classification for evaluation. The results showed that the DNN algorithm achieved an accuracy value of 99.76% and a loss value of 0.006%, outperforming the other algorithms. This study highlights the importance of implementing intrusion detection in protecting IoT devices and networks and demonstrates the efficacy of the DNN algorithm in detecting and preventing network security breaches.