Electronic Thesis and Dissertation Repository

Thesis Format

Monograph

Degree

Master of Science

Program

Computer Science

Supervisor

Haque, Anwar

Abstract

The advancement of internet technology and growing involvement in the cyber world have made us prone to cyber-attacks inducing severe damage to individuals and organizations, including financial loss, identity theft, and reputational damage. The rapid emergence and evolution of new networks and new opportunities for businesses and technologies are increasing threats to security vulnerabilities. Hence cyber-crime analysis is one of the wide range applications of Data Mining that can be eventually used to predict and detect crime. However, there are several constraints while analyzing cyber-attacks, which are yet to be resolved for more accurate cyber security inspection.

Although there are many strategies for intrusion detection, predicting upcoming cyber threats remains an open research challenge. Hence, this thesis seeks to utilize temporal correlations among attack frequencies within specific time periods to predict the future severity of cyber incidents. The research aims to address the current research limitations by introducing a real-time data collection framework that will provide up-to-date cyber-attack data. Furthermore, a platform for cyber-attack trend analysis has been developed using Power BI to provide insight into the current cyber-attack trend. A correlation was identified in the reported attack volume across consecutive time frames through collected attack data analysis. This thesis introduces a predictive model that forecasts the frequency of cyber-attacks within a specified time window, using solely a historical record of attack counts. The research includes various machine learning and deep learning methods to develop a prediction system based on multiple time frames with an over 15% improvement in accuracy compared to the conventional baseline model. Namely, our research demonstrates that cyber incidents are not entirely random, and by analyzing patterns and trends in past incidents, developed AI techniques can be used to improve cybersecurity measures and prevent future attacks.

Summary for Lay Audience

Increasing worldwide connectivity of networks has exacerbated the risk of cyber-attacks, and due to the nature of the networks, many organizations are prone to being robbed of their confidential data by cybercriminals. With the help of machine learning, cybersecurity systems can learn from their patterns and prevent attacks that can help cyber security professionals become more proactive in addressing threats.

This thesis develops a framework for real-time data collection to provide cyber-attack analysts with recent data and a platform for cyber-attack trend analysis to visualize the general patterns globally. Furthermore, the research employs various AI techniques to utilize temporal correlations between attack frequencies within specific time frames in order to predict the future severity of cyber incidents by predicting attack volume for a particular time window. Hence, it will also help cyber security analysts prepare for cyber security measures. Namely, our contribution focuses on developing a data collection process, presenting a visual representation of cyber-attack trends, and predicting different cyber-attack intensities by enhancing prediction accuracy to assist in analyzing future cyber-attack severity in order to prevent potential cyber-crimes.

Download

Share

COinS