Keylogger Malware Detection Using Machine Learning Model for Platform-Independent Devices

Keyloggers, which are programs created to record each keystroke performed on a computer, give the possibility to steal significant amount of critical information without the owner's consent. Malware-infected software is frequently used by online criminals to attack mobile devices like smartphones and tablets. But as time goes on, hackers get more intelligent. Adding a keylogger to a website rather than a software program will be simpler for them because the latter requires customers to download and install it on their devices before accessing it. The goal of this paper is to provide a machine learning-based framework for website keylogger detection mechanism for platform-independent devices. The security and privacy of Internet users would be improved by this technology. This study employs XGBoost, LightGBM, and CatBoost as classifiers. This study also uses an Exhaustive Feature Selection approach for feature selection which made the detection process more reliable and reduce runtime complexities. The simulation results show that XGBoost achieved the accuracy of 91% for full feature set and 98% of selected features which are better than other existing machine learning and boosting techniques.