A Data-Driven Classification Framework for Cybersecurity Breaches

Abstract

Unauthorized access to sensitive or confidential data results in a data breach, which can cause significant harm to an organization. Reporting breaches and reviewing prior records can help reduce damages. To aid in preparation, antivirus and security companies have published data breach reports, but they can be difficult to comprehend and require substantial effort to study. This article proposes a data breach incident classification framework using machine learning algorithms (naive Bayes, logistic regression, support vector machine, and random forest) on a dataset from the Privacy Rights Clearinghouse. The framework’s performance is evaluated using various metrics, including accuracy, F1 score, and confusion matrix. The article also employs topic modeling with latent Dirichlet allocation to enhance the classification’s accuracy.