Malware cyberattacks detection using a novel feature selection method based on a modified whale optimization algorithm

Malware cyberattacks have increased rapidly with the rise of Internet users, IoT devices, smart cities, etc. Attackers are constantly trying to evolve their methods and attack techniques to exploit human vulnerabilities and non-existing system vulnerabilities. In a malware attack, a user is tricked into giving personal information, such as login credentials or credit card information, to something that appears trustworthy. When this sensitive information falls into the hands of hackers, it serves as the basis for further malicious activity. In recent years, numerous researchers have proposed a machine learning-based strategy for detecting malware attacks; however, they have used many features to improve reliable malware detection approaches. Many malware detection methods require high computational power, so they cannot be used on devices with limited resources. We proposed a new system to detect malware attacks by feature selection based on a modified whale optimization algorithm to address these issues. An experimental benchmark dataset called ISCXURL-2016 is used to evaluate our system. The proposed system was tested against five machine learning classifiers, and it was found that XGBOOST had the highest accuracy of 99.66% and the lowest false positive rate of 0.23%.