A New Hybrid Approach using Deep Learning in handling IoT Attack

The term IoT, or Internet of Things, is revolutionizing the world today. It refers to all physical objects that can connect to the Internet. However, with the strong growth of many IoT devices and the growth of the IoT, new security threats are emerging due to vulnerabilities that can generate very critical impacts. In this work, we propose a network intrusion detection model using a deep learning approach, called Convolutional Auto- Encoders, to detect malicious activities in IoT networks. To evaluate our approach, we used the IoT-23 dataset. The results show that the convolutional autoencoder model gives good and promising predictive accuracy in detecting different types of attacks, with a high accuracy and recall rate, reaching an accuracy of 99.88%.