Improving Intrusion Detection in Internet of Things Networks with Feed-Forward Neural Networks Based on the UNSW-NB15 Dataset
Madhuri Telidevara, D. Kothandaraman · 2023
The research presented here proposes a new technique for identifying malicious activity in IoT networks by analyzing data from the UNSW-NB15 dataset. The considered approach uses a feed-forward neural network to detect malicious behaviour. One input layer, two hidden layers, and one output layer make up the feed-forward neural network's design. The hidden layers are used to detect malicious actions, whereas the input layer contains characteristics taken from the dataset. The output layer gives the verdict on the categorization. After testing, it was determined that the under consideration method achieved an accuracy of around 0.88 during attacks and around 1.0 during regular data sampling. According to the results, the proposed method is able to accurately detect malicious activity within IoT networks. The proposed technique provides a reliable and efficient approach for detecting malicious behaviors within IoT networks. In addition, the proposed method may be easily included into existing IoT infrastructures.