Malicious codes detection based on ensemble learning

摘要: As malicious codes become more complex and sophisticated, the scanning detection method is no longer able to detect various forms of viruses effectively. In this paper, we explore solutions based on multiple classifiers fusion not strictly dependent certain code. Motivated by standard signature-based technique for detecting viruses, idea automatically code using n-gram analysis. After selecting features information gain, probabilistic neural network used in process building testing proposed multi-classifiers system. Each one individual produce classification evidences. Then these evidences are combined Dempster-Shafer combination rules form final results new Experimental produced engine shows improvement compared classifiers.