| Autor: |
Chinchalkar, Soham P., Somkunwar, Rachna K. |
| Předmět: |
|
| Zdroj: |
Revue d'Intelligence Artificielle; Feb2024, Vol. 38 Issue 1, p269-275, 7p |
| Abstrakt: |
Every computer user deals with serious privacy and security challenges. Keyloggers are a type of software malware that records keystroke events from the console and saves them to a log file. It allows to obtain sensitive information like passwords, PINs, and usernames and communicates with vengeful attackers without attracting the attention of users. Keyloggers are also types of session hijackers that record user keystrokes made on the computer to steal any sensitive information from the system. Keyloggers are the most dangerous and covert malware for our system since they are difficult to detect because they run in the background of the computer. The primary issue with keylogger detection in a system is its time-consuming nature and its reliance on a particular type of input traffic behaviour. Keyloggers can be prevented using antiviruses, but, cannot be detected once they entered into the system. We proposed a system that combines Dendritic Cell Algorithms (DCA) and Machine Learning Algorithms (MLA) to address these problems. Our system can accurately detect a software keylogger if it is present which is based on the rate at which inputs are given to the system. The best accuracy was attained by our hybrid SVM-NB-DCA and SVM-DCA approach, with accuracies of 99.8% and 96%, respectively. Hence, results have shown that our hybrid system is effective and accurate for keylogger detection. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
