Risk Analysis of Device Within the Organization that are Vulnerable to Cyber Security Attacks with Artificial Intelligence
Tanupat Ngampunprasert, Mahasak Ketcham · 2024
This research presents an analysis of internal organizational device risks susceptible to cyber security attacks using artificial intelligence, specifically utilizing the Gradient Boosted Trees algorithm. The study utilizes data from Internet Traffic Logs generated by a firewall, recording incidents of cyber security attacks over the year 2021, consisting of 1,048,575 records with 45 columns. The data is categorized into 21 classes based on the type of cyber security attack. The dataset is divided into two sets: the first set, constituting 70% of the entire dataset, is used for training the model, while the second set, constituting 30%, serves as the testing dataset. Both sets are non-overlapping, and data preparation has been performed. The algorithm's performance was evaluated, revealing that the Gradient Boosted Trees algorithm achieved the highest accuracy at 94.64%. The analysis accurately predicted cyber security attacks, especially for 12 classes with accuracy exceeding 80%, while the remaining 9 classes had accuracy below 80%. The results of the analysis are visualized using Microsoft Power BI.