PEDM: Pre-Ensemble Decision Making for Malware Identification and Web Files

摘要: Connecting your system or device to an insecure network can create the possibility of infecting by unwanted files. Malware is every malicious code that has potential harm any computer network. So, detecting harmful files a crucial duty and important role in system. Machine learning approaches use variety features such as Opcodes, Bytecodes, System-calls achieve accurate malware identification. Each these feature sets provides unique semantic view, while, considering effect altogether more reliable detect attacks. disguise itself some views, but hiding all views will be much difficult. Multi-View Learning (MVL) outstanding approach considers multiple problem improve overall performance. In this paper, inspiring MVL proposed incorporate various exploit complementary information identify file. way, consensus used minimize error classifier based on sparse representation. To show generalization power method, datasets are employed. Experimental results indicate addition high performance, method advantage overcoming imbalanced conditions.