Malware detection has become a challenging task due to the increase in the number of malware families. Universal malware detection algorithms that can detect all the malware families are needed to make the whole process feasible. However, the more universal an algorithm is, the higher number of feature dimensions it needs to work with, and that inevitably causes the emerging problem of Curse of Dimensionality (CoD). Besides, it is also difficult to make this solution work due to the real-time behavior of malware analysis. In this paper, we address this problem and aim to propose a feature selection based malware detection algorithm using an evolutionary algorithm that is referred to as Artificial Bee Colony (ABC). The proposed algorithm enables researchers to decrease the feature dimension and as a result, boost the process of malware detection. The experimental results reveal that the proposed method outperforms the state-of-the-art.
翻译:由于恶意软件家庭数量的增加,发现恶意软件已成为一项艰巨的任务。需要通用的恶意软件检测算法来检测所有恶意软件家庭,才能使整个过程变得可行。然而,更普遍的算法是,它需要处理的特征层面数量较多,这不可避免地造成新的尺寸问题。此外,由于恶意软件分析的实时行为,也很难使这一解决方案发挥作用。在本文件中,我们处理这一问题,并打算提出基于特性选择的恶意软件检测算法,使用一种进化算法,称为人工蜂窝(ABC)。提议的算法使研究人员能够减少特性层面,从而推动恶意软件检测过程。实验结果表明,拟议方法超出了工艺的状态。