Fast & Furious: On the modelling of malware detection as an evolving data stream

作者：

Highlights：

• Improved version of Drebin dataset containing the release date of each app and a representative version of AndroZoo dataset.

• Novel data stream cycle that considers changes in classifier and feature extractor.

• Experiments showing improvement in performance in comparison to existing approaches, such as DroidEvolver.

• In-depth analysis of the vocabulary used by a feature with cybersecurity background.

摘要

•Improved version of Drebin dataset containing the release date of each app and a representative version of AndroZoo dataset.•Novel data stream cycle that considers changes in classifier and feature extractor.•Experiments showing improvement in performance in comparison to existing approaches, such as DroidEvolver.•In-depth analysis of the vocabulary used by a feature with cybersecurity background.

论文关键词：Machine learning,Data streams,Concept drift,Malware detection,Android

论文评审过程：Received 9 February 2021, Revised 13 August 2022, Accepted 13 August 2022, Available online 22 August 2022, Version of Record 9 September 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118590