Avaliação de métodos de classificação baseados em regras de associação para detecção de malwares android

Abstract

This research aims to investigate the performance and viability of different models of association rules in the context of classifying Android malwares. For that, we developed a new classification model based on association rules and rule quality. For model comparison purposes, we used known and frequently used datasets for training Android malware detection models. The results demonstrate that our model performs equivalently to other models based on association rules, obtaining accuracy values above 85\%, and in some cases surpassing machine learning models.