An approach towards hybrid feature selection for detection of DDoS attack. (14th April 2021)

• Record Type: Journal Article
• Authors: Patil, Anuja; Kshirsagar, Deepak
• Is Part Of: International journal of autonomic computing. Volume 3:Number 3/4(2020)
• Page Start: 274
• Page End: 289
• Publication Date: 2021-04-14
• Subjects: feature selection -- information gain -- DDoS attack; Autonomic computing -- Periodicals; Computer science -- Periodicals; 004.05
• Languages: English
• ISSNs: 1741-8569

Abstract

Even though the organisation uses various security measures the attacks over the internet are increasing day by day. This paper proposes a hybrid feature selection model for the detection of a DDoS attack. In this paper, the two-step hybrid feature selection method is used. The CICIDS2017 dataset with 84 features is used for the implementation. Information gain, gain ratio, and correlation filter-based algorithms are used for the ranking of features and then the forward selection approach is used to reduce the features up to 32. The system gives higher accuracy of 88.7373% for the correctly classifying DDoS attack.