Identification of malware activities with rules

The article describes the method of malware activities identification using ontology and rules. The method supports detection of malware at host level by observing its behavior. It sifts through hundred thousands of regular events and allows to identify suspicious ones. They are then passed on to the second building block responsible for malware tracking and matching stored models with observed malicious actions. The presented method was implemented and verified in the infected computer environment. As opposed to signature-based antivirus mechanisms it allows to detect malware the code of which has been obfuscated.