From Data to Defense: Optimizing Cyber Resilience by AI-Driven Vulnerability Prioritization

<p>In the current digital era, cybersecurity has emerged as a major responsibility for companies everywhere. Due to more sophisticated cyber-attacks, IT systems are becoming more complicated. Thus, the effective vulnerability management solutions are becoming more and more important. Prioritizing risks is important since it helps businesses allocate resources and deal with the most serious security concerns. An overview of vulnerability prioritizing techniques is provided in this document, with a focus on the importance of precisely assessing and ranking vulnerabilities according to their base score and the title of the risk. A formula has been proposed by assigning weights for the base score and the title. By taking a variety of base-to-title ratios, we achieved accurate results for 7:3 ratio. By using this ratio, we prioritized the threats and classified them based on achieved priority score. The classification task is done for the self-prepared dataset in which we used five different algorithms. It includes, SVM, Naïve Bayes, Neural Network, XG Boost, Gradient Boosting. Out of all, XG Boost algorithm performed well with an accuracy of 96.7 percent. By using this approach, organizations can rank their threats and allocate them to the resources effectively.</p>

Name thesis:degree_name

Master of Science (MS)

Level thesis:degree_level

Masters Thesis

Year dc:date.available

2024

Author dc:creator

• Penchala, Sindhuja

Contributors dc:contributor

• Dr. Nick Rahimi
• Dr. Andrew H Sung
• Dr. Partha Sengupta

• Vulnerability Prioritization
• Cyber Security
• Threat Detection
• Classification
• XG Boost
• Computational Engineering
• Computer Engineering
• Education
• Educational Assessment, Evaluation, and Research
• Risk Analysis

Repository record dc:identifier

https://aquila.usm.edu/masters_theses/1078

OAI identifier oai:identifier

oai:aquila.usm.edu:masters_theses-2163