Evaluating Machine Learning Algorithms to Detect and Classify DDoS Attacks in IoT

A Distributed Denial of Service (DDoS) attack is a lethal threat to web-based services and applications. These attacks can cripple down these services in no time and deny legitimate users from using these services. The problem has further prevailed with the massive usage of unsecured Internet of Things (IoT) devices across the Internet. Moreover, many existing rule-based detection systems are easily vulnerable to attacks. In this paper, we performed a comparative analysis of Machine Learning (ML) algorithms to detect and classify DDoS attacks. As part of the work, various machine learning algorithms such as Naïve Bayes, J48, Random Forest and ZeroR ML classifiers are compared. Principal Component Analysis (PCA) method has been used to select the optimal number of features. WEKA tool has been used to implement ML algorithms.