Cyber Threat Detection on Internet of Things

The internet of things (IoT) is a connected network of devices that has the ability to communicate with each other and bring data to users using internet. The amount of collected data from loT in different domain is so huge and conventional analysis of these data is very hard and need more feasible methods. There are huge calls in loT for appropriate security and privacy policies to prevent potential cyber threats. One of the security requirements of the loT user is the confidentiality and integrity of the data collected by the sensors. Recently deep learning is prominent approach to combat intrusion in loT and several deep learning models have been proposed for it. However, the field is still evolving and there is ongoing research to further improve the performance of the system. We have selected deep learning algorithms like LSTM, AUC, GRU, RNN and MLP and compared with other state of art methods like CNN, SVM and others. The models RNN, LSTM and GRU, an Accuracy of 100%, recall of 100%, and precision of 100% were recorded. Compared with AUC and MLP, the RNN, LSTM and GRU records higher in accuracy, recall and precision.