Improving JavaScript Malware Classifier's Security against Evasion by Particle Swarm Optimization

Zibo Yi, Jun Ma, Lei Luo, Jie Yu, Qingbo Wu

Abstract: Machine learning techniques have recently been applied to JavaScript malware detection. However, the detection can be misled since a malicious script may be modified by an adversary then masqueraded as a benign one. In this paper, we investigate how these evasion attacks work and propose a metric to measure the classifier's security against it. To improve the security, we propose a feature selection approach using particle swarm optimization. The experiments validate that our approach can strengthen classifier's security with its accuracy also increases.