Detecting IoT Attacks Using an Ensemble Deep Learning Model

As we know, Gartner identified cloud computing as one of the top ten strategic technologies for ۲۰۲۰. On the other hand, over the past few decades, it has gained immense popularity due to its ease of use and flexibility, with various types on the rise. However, there are numerous challenges in ensuring better IoT services, such as complexity, compliance, security, privacy, and more. Security in cloud computing is a critical issue, and researchers in this field consider it a primary concern. In January ۲۰۲۰, Gartner identified security as one of the four influencing factors in IoT. The appeal and multiple advantages of access for users who are vulnerable to cyber threats increase cybercriminal interest in IoT. Therefore, ensuring security and privacy in the cloud and IoT space is vital. Given the importance of security in IoT and the solutions that various AI and machine learning methods provide for identifying cyberattacks, we utilized an innovative approach that combines deep learning methods for detecting attacks in IoT environments. In this proposed model, we harnessed the power of three deep learning models: convolutional neural networks, long short-term memory, and gated recurrent units to enhance performance, combining them, and this hybrid model was evaluated on the KDD'۹۹ open-source dataset. The proposed method, with an accuracy of ۹۷.۹%, precision of ۹۸.۸%, and a true negative rate of ۹۸.۹%, shows that it outperforms other machine learning and deep learning methods. In other words, the combination of these methods has led to more consistent accuracy in detecting cyber attacks.

کلیدواژه ها:

Network security ، Attacks Detection ، Internet of Things (IoT) ، Deep Learning ، Ensemble Learning