Evaluating the Impact of Binary Code Feature Selection on Explainable AI Models for Ransomware Detection

The increasing prevalence of ransomware attacks poses a significant challenge to cybersecurity, necessitating the development of sophisticated detection methods that are both accurate and interpretable. Incorporating binary code feature selection into Explainable AI (XAI) models presents a novel approach that enhances the models' ability to identify critical features, thereby improving classification performance and transparency. This article provides a comparative analysis of XAI models with and without the integration of binary code feature selection, demonstrating substantial improvements in accuracy, interpretability, and computational efficiency when the feature selection technique is applied. The findings indicate that the refined feature sets not only lead to more accurate predictions but also contribute to a clearer understanding of the decision-making process, which is crucial for the trustworthiness and applicability of the models in real-world scenarios. The study's results highlight the potential of feature selection to advance ransomware detection systems, contributing to the development of more reliable and effective cybersecurity solutions.