Smart Media Journal (스마트미디어저널)

THE KOREAN INSTITUTE OF SMART MEDIA (한국스마트미디어학회)
권호 표지
DOI QR코드

DOI QR Code

A Review of Public Datasets for Keystroke-based Behavior Analysis

  • Kolmogortseva Karina (Department of Artificial Intelligence Convergence, Chonnam National University) ;
  • Soo-Hyung Kim (School of Artificial Intelligence, Chonnam National University) ;
  • Aera Kim (AI Convergence Department as a research staff)
  • Received : 2024.06.17
  • Accepted : 2024.08.05
  • Published : 2024.07.31
Download PDF
⟨ Previous Next ⟩

Abstract

One of the newest trends in AI is emotion recognition utilizing keystroke dynamics, which leverages biometric data to identify users and assess emotional states. This work offers a comparison of four datasets that are frequently used to research keystroke dynamics: BB-MAS, Buffalo, Clarkson II, and CMU. The datasets contain different types of data, both behavioral and physiological biometric data that was gathered in a range of environments, from controlled labs to real work environments. Considering the benefits and drawbacks of each dataset, paying particular attention to how well it can be used for tasks like emotion recognition and behavioral analysis. Our findings demonstrate how user attributes, task circumstances, and ambient elements affect typing behavior. This comparative analysis aims to guide future research and development of applications for emotion detection and biometrics, emphasizing the importance of collecting diverse data and the possibility of integrating keystroke dynamics with other biometric measurements.

Keywords

Acknowledgement

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (RS- 2023-0219107), and the Institute of Information & communications Technology Planning & Evaluation (ITP) under the Artificial Intelligence Convergence Innovation Human Resources Development (ITP-2023-RS-2023-025629) grant funded by the Korea government(MSIT).

References

  1. C. Epp, M. Lippold, Mandryk, "Identifying Emotional States Using Keystroke Dynamics", Proceedings of the 2011 Annual Conference on Human Factors in Computing Systems (CHI 2011), pp. 715-724, 2011.
  2. Amith K. Belman, Li Wang, Sundaraja S. Iyengar, Pawel Sniatala, Robert Wright, Robert Dora, Jacob Baldwin, Zhanpeng Jin, Vir V. Phoha, November 20, 2019, "SU-AIS BB-MAS (Syracuse University and Assured Information Security - Behavioral Biometrics Multi-device and multi-Activity data from Same users) Dataset ", IEEE Dataport, doi: https://dx.doi.org/10.21227/rpaz-0h66.
  3. Sun, Y.; Hayreddin, C.; Shambhu, U. "Shared Keystroke Dataset for Continuous Authentication," In 2016 IEEE International Workshop on Information Forensics and Security (WIFS); IEEE: Abu Dhabi, United Arab Emirates, 2016.
  4. Christopher Murphy, Jiaju Huang, Daqing Hou, Stephanie Schuckers. "Shared Dataset on Natural Human-Computer Interaction to Support Continuous Authentication Research," IJCB 2017, Colorado Denver, 2018.
  5. K. S. Killourhy and R. A. Maxion, "Comparing anomaly-detection algorithms for keystroke dynamics," in 2009 IEEE/IFIP International Conference on Dependable Systems & Networks, pp. 125-134, 2009.
  6. Maalej, Aicha, and Ilhem Kallel. "Does keystroke dynamics tell us about emotions? A systematic literature review and dataset construction," 2020 16th International Conference on Intelligent Environments (IE). IEEE, Aug., 2020.
  7. A. Kolakowska, ''Recognizing emotions on the basis of keystroke dynamics,'' in Proc. 8th Int. Conf. Hum. Syst. Interact. (HSI), Jun., pp. 291-297, 2015.
  8. Kolakowska, A. and Landowska, A., 2021. "Keystroke Dynamics Patterns While Writing Positive and Negative Opinions," Sensors, vol. 21, no. 17, pp. 5963, 2021.