Artificial Intelligence based Model Using Various Machine Learning Techniques to detect DDoS Attack

The magnanimous increase in IoT devices has led to the attention of hackers and malicious user at a massive rate, with the increase in devices and their applications the attacks on the devices are also increasing. These attacks aim to disrupt the quality of services and can be fatal sometimes, the common attack of all is the Denial-of-service (DoS) attack. DoS attack targets the device by exhausting the system by using its all resources and therefore denying the service to the end-users. This paper discusses both Unsupervised and Deep Learning (DL) techniques to detect DoS attacks in IoT devices. We have used Knowledge Discovery in Databases (KDD) unlabelled datasets to achieve our goal, clustering, and deep auto-encoding techniques were used for classifying and training the data, and finally Supervised learning was used for the evaluation of the classifiers in detecting the DoS attacks. The finding from the studies claims that most of the classifiers seem to be performing well in detecting the DoS attack except Naive Bayes which seems to be performing on average, also the various clustering techniques provides encouraging results except the Gaussian mixture model which seems to be performing average in all the cases as compared to other clustering techniques.