We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
An Enhanced Intrusion Detection System for Protecting HTTP Services from Attacks.
- Authors
Al-Mimi, Hani; Hamad, Nesreen A.; Abualhaj, Mosleh M.; Daoud, Mohammad Sh.; Al-dahoud, Ali; Rasmi, Mohammad
- Abstract
Cybercriminals constantly develop new sophisticated methods to breach the security of their targets, thereby increasing cyberattack cases. HTTP is one of the earliest and most susceptible network services. However, it has several security flaws that have been exploited repeatedly throughout time. Constructing a robust intrusion detection system that prevents unauthorized access to network resources is crucial to recognize HTTP attacks and protecting data. Several novel ways have been recently proposed as a panacea for intrusion detection. However, building a secure HTTP system remains challenging because attackers continuously adapt their strategies to circumvent the system's security features. The current study employed machine learning (ML) classifiers to build a model that categorizes the HTTP traffic as an attack or normal traffic. The proposed model is called ML-HTTP. The UNSW_NB15 dataset was utilized to evaluate the performance of the ML-HTTP model. Among several classifiers investigated with the ML-HTTP model, both random forest and logistic regression classifiers achieved the highest accuracy (96.32%), recall (97.01%), precision (97.63%), and Matthew's correlation coefficient (91.86%).
- Subjects
HTTP (Computer network protocol); MACHINE learning; RANDOM forest algorithms; LOGISTIC regression analysis; CYBERTERRORISM; STATISTICAL correlation
- Publication
International Journal of Advances in Soft Computing & Its Applications, 2023, Vol 15, Issue 2, p67
- ISSN
2710-1274
- Publication type
Article
- DOI
10.15849/IJASCA.230720.05