A Comparative Study Of Binary Class Logistic Regression and Shallow Neural Network For DDoS Attack Prediction

• Published in: Proceedings of IEEE Southeastcon pp. 310 - 315
• Main Authors: Tufail, Shahid, Batool, Shanzeh, Sarwat, Arif I.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 26.03.2022
• Subjects: Analytical models; Cyber Attacks; Data Packets; DDOS Attacks; Denial-of-service attack; Distributed Denial of Service; DOS Attacks; Logistic Regression; Machine learning; Machine Learning Algorithm; Multi Layer Perceptron; Networks; Neural networks; Predictive models; Software algorithms; Training
• ISSN: 1558-058X
• DOI: 10.1109/SoutheastCon48659.2022.9764108

In the area of internet security, cybersecurity is a serious subject. Every industry is witnessing thousands of cyberattacks every year. Among the most deadly cyber-attacks are the distributed denial-of-service attack (DDOS) and the False data injection attack (FDIA). In this paper, we performed a comparative study for predicting DDOS attacks using two machine learning algorithms that are logistic regression and shallow neural network(SNN). In logistic regression, we achieved an accuracy of 98.63% and for SNN accuracy we achieved was 99.85%. However, our study shows that the training time was exponentially higher for SNN in comparison to logistic regression.