UNECE Threat List Case Study: Prediction of Cyber Risks in the Automotive Domain Using Regression Analysis

Abstract

AbstractCybersecurity has become a significant factor in the automotive industry. It is responsible for protecting data and critical components of the vehicle against cyberattacks. Automobile manufacturers should ensure their vehicle’s cybersecurity, considering the UNECE regulation for road vehicle approval. Estimating risk severity is critical in this domain for specifying appropriate security mechanisms to address existing cyber risks. To have an accurate risk evaluation, we need to consider multiple factors of likelihood and impact, which indicate the probability of risk occurrence and its severity. However, establishing a relationship amongst multiple factors of impact and likelihood remains a challenging problem that needs to be addressed. In this work, we propose applying machine learning regression analysis to create correlations between various independent factors of likelihood and impact for predicting dependent variables that indicate the assessment of cyber risk severities. We use the UNECE threat list as a case study to show how machine learning regression approaches may help predict realistic cyber risk estimations in the automotive domain. Finally, we evaluate outcomes to demonstrate the effectiveness of regression analysis for evaluating cyber risks in the automotive sector.KeywordsCybersecurityThreatsRisk evaluationRegression analysis