| Makhamadaziz Rasulmukhamedov, Gulmira Mirzaeva, Nomaz Mirzaev, Nuraddin Gafforov Extraction of Features Characterizing Keystroke Dynamics in the User Identification problem |
|---|
| Abstract. This article examines the separation of characteristics describing the user's behavior in the matter of identifying the user based on keyboard dynamics. The study analyzed time sequences recorded by users while working with the keyboard. Haar wavelet transformation was used to distinguish characteristics that characterize user properties. The obtained wavelet coefficients were used as characters characterizing users. The advantages of the highlighted features were checked using SVM, MLP, Random Forest, and KNN algorithms. The results obtained can be used in the development of biometric authentication systems. |
| Keywords: Keyboard dynamics, user identification, Haar wavelet transform, SVM, MLP, KNN, Random Forest, time series, hold time, flight time |
 Download PDF |
| DOI: https://doi.org/10.54381/itta2026.1.09 |