Abstract:

Online education is changing the world-wide education scenario (e.g. Massive Open Online Courses and Participatory Open Online Courses). Online courses break with the barriers associated to traditional lessons and provide a new widely accessible education over the internet. A new education emerges on the basis of open, free, distributed and participatory courses. This new approach does not attempt to replace the traditional education and it must be seen as a complementary proposal to overcome limitations and take advantage of new technologies. However, there is a discussion among all the academic sectors about this new educational scenario. Among the topics under discussion, how to certificate the successful completion of courses is one of the most controversial. How can we certify courses without classroom attendance? Is it possible a reliable authentication of the students? How can we avoid/detect fraudulent users? The widely and accessible nature of online courses complicates the individualization of the education and student authentication is part of this individualization. The aim is not to criminalize the students but reliable authentication is critical for the future of online education and solutions should be balanced between security and user's privacy. Therefore, researchers and instructors have made efforts to propose authentication approaches which comply with the characteristics of these courses. Among all the technologies proposed, biometric authentication is one of the most attractive approaches. Biometric recognition technologies allow to authenticate users based on "something that we are" instead of traditional authentication based on "something that we know" such as passwords or PINs.

In this work we present:
i) an analysis of biometric technologies for online student authentication.
ii) a case study on keystroke dynamics authentication applied to a real operational environment.

Concisely, this work studies different biometric technologies and their advantages/disadvantages for online student authentication services. The analysis is made on the basis of three main pillars: performance, security and user acceptance. Our study provides new insights on the deployment of these technologies in educational environments with special attention to keystroke dynamics systems. Keystroke dynamic authentication has attracted much interest from industry and researchers during the last decade.

Keystroke dynamics authentication is interesting for online authentication because it is:
i) transparent (it runs in the background without requiring explicit user interaction).
ii) continuous (authentication can be performed over all the user activity, not only based on an initial access to the platform).

This work includes a case study on keystroke dynamic authentication over the typing patterns of 64 students answering questions in 3 online exams over a semester. We analyze the performance of four keystroke biometric systems and the results obtained show a promising performance with a correct student authentication over 90%. This work encourages to further explore in authentication services based on biometric technologies for online courses.

Keywords:

Innovation, technology, authentication, keystroke dynamics, biometrics.