A Comparative Study on Feature Selection Method for N-gram Mobile Malware Detection

Mohd Zaki Mas'ud ； Shahrin Sahib ； Mohd Faizal Abdollah ； Siti Rahayu Selamat ； Choo Yun Huoy

《International Journal of Network Security》 19卷5期 (2017/09) Pp. 727-733

https://doi.org/10.6633/IJNS.201709.19(5).10

摘要

In recent years, mobile device technology has become an important necessity in our community at large. The ability of the mobile technology today has become more similar to its desktop environment. Despite the advancement of the mobile devices technology provide, it has also ex- poses the mobile devices to the similar threat it predecessor possess. One of the anomaly based detection methods used in detecting mobile malware is the n-gram system call sequence. However, with the limited storage, memory and CPU processing power, mobile devices that provide this approach can exhaust the mobile device re- sources. This is due to the huge amount of system call to be collected and processed for the detection approach. To overcome the issues, this paper investigates the use of several different feature selection methods in optimizing the n-gram system call sequence feature in classifying benign and malicious mobile application. Several filter and wrapper feature selection methods are selected and their performance analyzed. The feature selection methods are evaluated based on the number of feature selected and the contribution it made to improve the True Positive Rate (TPR), False Positive Rate (FPR) and Accuracy of the Linear-SVM classier in classifying benign and malicious mobile malware application.

關鍵字

Feature Selection ； Linear SVM ； Mobile Mal-ware ； Mobile Malware Detection ； N-gram