Evaluation and detection of cybercriminal attack type using machine learning

Nakid, Safvaan Shadab (2021) Evaluation and detection of cybercriminal attack type using machine learning. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (859kB) | Preview
Preview	PDF (Configuration manual) Download (424kB) | Preview

Abstract

There has been a significant rise in data breaches and various types of computer-based attacks that result in monetary as well as infrastructural losses to both individuals and organizations alike. Law enforcement agencies and organizations often find themselves at crossroads at such times whilst dealing with unprecedented attacks or breaches. As cyber offenders or perpetrators evolve in their attack patterns rather rapidly than the organizations equipped with defensive mechanisms, it becomes virtually difficult to trace back to the attack patterns of the criminal. With its negative effects such as breach of confidentiality and integrity, data sustained by the organizations such as event logs, reports can also be used to gain further insights as to how a criminal can potentially harm a system & understand what vulnerable areas of the system infrastructure can be further identified to strengthen them.

This research paper addresses the issue of identifying such attack patterns and presents a model based on feature selection to understand the type of attack pattern employed by the offender which would then help narrow down the approach of creating the profile of the offender. A dataset of 1145 recorded data breaches and ransomware attacks from the University of Queensland was used in the research. As the dataset consisted of imbalanced columns, ROS and RUS sampling techniques were employed along with data tuning procedures and label encoding. Classification models such as Random Forest, KNN, Logistic regression were implemented on the data to identify the accurate attack-type of a given attack. Upon comparison of results, it was noted that Random Forest was able to outperform other models by achieving 95% accuracy with an average F1 score of 0.94.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QA Mathematics > Computer software > Computer Security T Technology > T Technology (General) > Information Technology > Computer software > Computer Security Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Cyber Security
Depositing User:	Tamara Malone
Date Deposited:	22 Dec 2022 13:34
Last Modified:	07 Mar 2023 12:49
URI:	https://norma.ncirl.ie/id/eprint/6031

Actions (login required)

View Item