Feature Extraction and Classification of Application Layer DDoS Attacks using Machine Learning Models
Sarabjeet Kaur, Amanpreet Kaur Sandhu, Abhinav Bhandari · 2023
Increased internet use has connected organizations to a shared network. These networks are vulnerable to many security threats and are victims of attackers. DDoS (Distributed Daniel of Service) attack exhausts network resources. These attacks target the network layer, transport layer and application layer. Though, it leaves the huge impact on resources at different levels. Especially in the application layer, DDoS attacks make the network inaccessible to legitimate users. It is crucial to detect application layer DDoS attack in network. Feature extraction is found helpful in getting high accuracy of detection through machine learning classifiers in this case. In this study, we have implemented efficient machine learning classifiers to detect application layer DDoS attacks using SDN dataset features. We have used Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), and Independent Component Analysis (ICA) to reduce the feature set of data. In addition to, machine learning classifiers are trained with extracted features and prediction of application layer DDoS attack is done. It is observed that LDA model extracted 1 feature out of 13 and gives maximum accuracy of detection for used classifiers. The result analysis of this study with Decision Tree, Random Forest, and Support Vector classifiers is achieved up to 99.6% and the proposed work is compared to previous studies to analyze the results.