A Hybrid Deep Learning Model for Attack Detection and Classification in IoT Network

The advent of Internet of Things (IoT) devices has introduced a wide array of security challenges. We find that which security breaches faced in IoT systems’ and what class of them they fall into is a priority in securing private information. In this paper we propose a hybrid deep learning model for the detection and classification of security threats in the Internet of Things. We combine the power of Random Forest with CNN Features, Decision Tree and Extreme machine learning to produce a model which is very good at analysing and classifying security threats. It takes into consideration features in the raw IoT data, and a few of the features in our model considers time and trend aspects in the data. For the training and evaluation of the model we put together a large set of our own simulated security attack scenarios in IoT networks. The results show that the model performs very well on the tasks proposed, thus showing givenness of the system for real time security supervision. Our model which utilises deep learning technologies also answers the deficiencies there are in the security of the Internet of Things. With the hybrid model proposed security personnel can quickly detect and reply to damaging intrusions which could affect the Internet of Things and important information. Future work could analyze the model performance, scalability and the diverse ecosystems of the Internet of Things.