Classification Performance Improvement of Keystroke Data
Mehmet Erdal Özbek · 2019 Innovations in Intelligent Systems and Applications Conference (ASYU) · 2019
In this work, the classification of keystroke data, a behavioral biometrics modality, is investigated. The users degrading the overall classification performance are identified by classifying two different databases composed of different number of individuals and passphrases. The improvement in performance is demonstrated by eliminating those users in the training phase of the classifier.