Novel active learning methods for enhanced PC malware detection in windows OS

作者：

Highlights：

• The challenge of malware signature update is formalized as an active learning task.

• We present and compare several active learning (AL) strategies.

• The best results are achieved using our AL method called Exploitation.

• With our AL methods the number of malwares acquired daily is increased substantially.

• AL methods improve the predictive performance of malware detectors.

摘要

•The challenge of malware signature update is formalized as an active learning task.•We present and compare several active learning (AL) strategies.•The best results are achieved using our AL method called Exploitation.•With our AL methods the number of malwares acquired daily is increased substantially.•AL methods improve the predictive performance of malware detectors.

论文关键词：Malware,Malicious code,Machine Learning,Active learning,SVM

论文评审过程：Available online 19 March 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.02.053