Akademik Veri Yönetim Sistemi
14th International Workshop on Computer Science and Engineering, WCSE 2024, Phuket Island, Tayland, 19 - 21 Haziran 2024, ss.269-278, (Tam Metin Bildiri)
One of the most important developments in authentication methods is the increasing popularity of user identification based on keystroke behaviors. To identify and verify people, this paper explores the topic of collecting and analyzing user-specific keystroke information. A complete dataset is generated by amassing unique characteristics including Hold, Down-Down, Up-Down keystroke durations, average number of keystrokes, backspace frequency, capitalization indicators, and shift key preferences. In this study, various machine learning pipelines are used to analyze the data set that contains these derived features and the associated user labels. Using these machine learning pipelines to correctly identify and attribute keyboard activities to their respective users is the main purpose. The goal of this work is to help build better user identification systems by using the nuances of keystroke properties. Results from experiments on the dataset show that the proposed machine learning pipelines successfully categorize and assign users according to their unique keyboard behaviors.