Cybersecurity: A Deep Learning Model for Intrusion Detection in IoT

Cybersecurity involves protecting the IoT devices from hackers, viruses, and malwares. This paper aims at building a robust deep learning-based IDS model to encounter cyberattacks. CNN-based IDS model is built on two datasets in which one is modified and other is raw dataset and their accuracies are compared to examine the impact of training datasets used in models. A comparative study of datasets used in previous researches have also been performed. After reviewing previous research papers built on trending datasets, it was found that most of the research lacks quality of dataset used in them. A CNN-based IDS model was built and trained on most recent publicly available CSE-CIC-IDS 2018 dataset and got an accuracy of 92%. The same model was trained on modified CSE-CIC-IDS 2018 dataset, the model achieved an accuracy of 94.67% which is better than the accuracy of the model of earlier not modified dataset. Thus, it proves that through proper data refining and extraction techniques, NIDS model used for cyberattacks can be made more precise and accurate.