Android malware detection method based on permission sequential pattern mining algorithm

Abstract

The permissions requested by Android applications reflect the behavior sequence of the application. While a generation of malicious behavior usually requires the cooperation of multiple permissions, so mining the association be-tween permissions can effectively detect unknown malicious applications. Most researchers concerned the statistical properties of a single permission, and there was little researchers studying the statistical properties of the association be-tween permissions. In order to detect unknown Android malwares, an Android malware detection method based on per-mission sequential pattern mining algorithm was proposed. The proposed method design a permission sequential pattern mining algorithm PApriori to dig out permissions association. PApriori algorithm could discover permission sequential pattern from 49 malware families and build the permissions association dataset to detect malware. The experiment results prove that it performs better than other related work in efficiency and accuracy.

Keywords

• sequential pattern mining;data mining;malware detection;permission feature;Android OS