Year: 2025
Type: article
Abstract: The expansion of digital infrastructure and the growing sophistication of cyberattacks expose limitations of traditional security controls. This paper examines AI-based anomaly detection using the CICIDS2017 dataset as a case study. Data preprocessing, feature selection, and application of Random Forest, SVM and Deep Neural Network models are described. Results synthesized from reproductions and benchmark studies indicate that AI approaches significantly improve detection accuracy and response time, while challenges remain in dataset quality, class imbalance and adversarial robustness .
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4414238335
FWCI:
Cites:
Cited by:
Related to: