Use of Classification Techniques to Predict Targets of Cyber Attacks for Improving Cyber Situational Awareness During the COVID-19 Pandemic

Abstract

As the world increasingly relies on online services, the risk and impact of cyber attacks also increases. In the arms race between cyber attackers and defenders, cyber security professionals need as much information as they can gather. Cyber situational awareness (CSA) is a broad strategy that aims to improve decision making in cyber security by analysing security events. This study aims to improve CSA by comparing data mining techniques, specifically classification techniques, when applied to cyber security data. The predictors are trained by classification algorithms and the training data is collected from Open Source Intelligence including cyber-attacks in Europe over the period 2017–2019. Furthermore, the techniques have been applied to data from a more recent period, during the COVID-19 pandemic in Europe. This has allowed the study to look at how COVID may have affected methods and targets of cyber attacks, and has shown a decrease in accuracy suggesting attack patterns have changed.