Canadian Journal of Pure & Applied Sciences

Abstract: Antivirus software developers are advocating for sophisticated antivirus designs to implement their antivirus systems. However, the current antivirus systems heavily rely on updating of their malicious signature databases to detect malicious codes in executable programs. The problem with frequent update of malicious signatures databases is that it is not scalable; it cannot detect malicious code whose signature is not in the malicious signature database. Consequently, we designed a generic antivirus system that does not contain malicious database but rather, malicious codes are detected by the type of operating system functions used by the executable program. The proposed generic antivirus system uses deterministic finite automata, Naïve Bayes and Chi square techniques to detect malicious codes in executable programs. When the generic antivirus system is deployed to any operating system environment, malicious codes can be accurately detected in executable programs without a need to update its malicious signature database.