Redefining Annual Rate Occurrence (ARO) in Risk Calculation with Probability and Regression

This study converts archaic Annual Rate of Occurrence(ARO) into a novel probabilistic model with the help of a public dataset while improving Annualized Loss Expectancy (ALE). It considers internal and external influence when calculating cybersecurity risk, where granularity is its key characteristic. In addition, the exact dataset is modelled by machine learning classifiers to identify types of breaches. Though the results were less than satisfactory, it provides a framework for future studies.