Page 26 - bookofabstract_vj5_final(1)

Android Malware Detection Using Machine Learning on Image Patterns

Fauzi Mohd Darus,Noor Azurati Ahmad @ Salleh & Aswami Fadillah Mohd Ariffin

Abstract

Android platform has been targeted by cyber-criminals due to the increase number of
Android users in 2017. More than 8,000 Android malware were identified everyday
making it is difficult for the malware analyst to detect them. Traditional malware
detection techniques are no longer reliable to detect newly created malware in short
period of time. In this paper, we use a different approach to detect Android malware.
The Android malware will be visualised into

gray scale images and their image features will be extracted using GIST descriptor. The
detection will be done and compare using five different classifiers namely Stochastic
Gradient Descent (SGD), k-nearest neighbor (kNN), Random Forest (RF), Support
Vector Machines (SVM) and Decision Tree (DT).

22