Avaliação de Abordagens para Classificação de Malware Bancário sob a presença de Concept Drift

Financial malware is an increasing threat due to the potential ofprofit for cybercriminals. Year after year, the arms race betweenmalware developers and security professionals foster the release ofmillions of malware variants. Although machine learning techniqueshave been successfully applied to malware classification tasks,the occurrence of concept drifts requires constant updates (or evenretraining) of these models. In this work, we evaluate how distinctmachine learning training approaches for malware classificationwhen we consider the arrival flow of samples as a data stream. Wealso observe the manifestation of concept drift in an actual datasetcomprised of thousands of financial malware samples, discussingdifferent incremental learning approaches to deal with a customdataset and how concept drift can arise in malware data.