IoT Network Intrusion Detection Using Long Short-Term Memory Recurrent Neural Network

Abstract

The Internet of Things (IoT) is the latest technologies for everyday physical devices which the digitally connected to the internet or each other. With the rapidly development of IoT platform, IoT have been encountered many malicious activities. IoT security is one of the most critical issues in developing and implementing of IoT platform. Intrusion Detection System (IDS) play an important role for security solution in IoT network. In our paper, we proposed an effective IDS model for intrusion detection in IoT network by using Long Short-Term Memory Recurrent Neural Network (LSTM RNN). Performance of proposed model to identify correctly the normal and attack has been evaluated on the benchmark intrusion dataset, UNSW-NB15 dataset which applied in the most of IoT intrusion detection research. Moreover, in this survey, we studied the performance of the proposed LSTM RNN IDS model by contrasting the Recurrent Neural Network (simple RNN) algorithm. The experimental results show the efficiency of proposed model with accuracy, precision, recall and F1 score. Our proposed LSTM RNN model outperformed than the simple RNN model to build a highly effective intrusion detection model with accuracy over 99% for IoT network attacks.