| DOI LINK: https://doi.org/10.59879/19PdW |
| Paper ID:19PdW |
| Volume:37 |
| Issue:12 |
| Title:Enhancing Cyber Threat Categorization with Artificial Intelligence: A Novel Clustering-Based Classification Framework |
| Abstract:This study investigates the application of artificial intelligence (AI) to improve cyber threat classification using clustering techniques. Leveraging the NF-UNSW-NB15-v2 dataset, the research addresses challenges such as data imbalance and overlapping attack patterns. The methodology integrates dimensionality reduction via Principal Component Analysis (PCA) and clustering using KMeans, focusing on features like transport layer ports, DNS query types, and network throughput. The experiments highlight the clustering algorithm's ability to identify inherent patterns within attack categories, though difficulties persist in distinguishing closely related attack types. Despite the imbalanced dataset, clustering by attack type revealed significant insights, enhancing the nuanced analysis of cyber threats. Evaluation metrics, including the silhouette score, emphasize areas for refinement. The findings demonstrate the potential of AI-driven clustering to complement existing cybersecurity frameworks, offering a pathway for more effective intrusion detection systems. This research underscores the importance of combining clustering with additional techniques to improve classification accuracy, advancing the capability of AI in addressing evolving cybersecurity threats. |
| Keywords:cybersecurity, clustering, artificial intellicence, KMeans, cyber attack |
| Authors:Matei Vasile CAPILNAS*, Adriana Mihaela COROIU |
| Paper PDF Link: View full PDF |