Ensemble Learning for Malware Detection

One of the major threats to security on the Internet nowadays is Malware. In fact, malware is frequently the root cause of the majority of Internet issues, including phishing and APT (Advanced Persistent Threat) attacks [1]. New methods have been proposed for detecting and preventing the most recent generation of malware. Nevertheless, there hasn’t been much research done on how effectively ensemble approaches perform. In this paper, we have designed two approaches based on ensemble learning to detect malware. The first approach uses a stacking method to create a set of machine learning classifiers, whereas the second approach combines several deep learning algorithms. For comparison, six individual machine learning algorithms and two deep learning algorithms were used. A dataset of 42,797 malware API call sequences and 1,079 goodware API call sequences is used for the evaluation of the suggested methods. The results of the experiments validate the proposed methods’ performance.