Abstract
E-learning became attractive choices for academic institutions owing to their benefits, especially offering transparent and fast students' evaluation as well as innovative teaching methods. However, security issues related to e-learning have been raised by its stakeholders mainly faculty members, students, and administration. Authenticate examination takers during the electronic evaluation (e- evaluation), comprising assessment (E-assessment) and exam (E-exam), is a major challenge. In this paper, an advanced scheme is proposed to resolve this imperative challenge by introducing an efficient secure approach for supervising e-evaluation. The scheme collects information related to students and their behavior during the courses' activities and uses them for controlling unethical behavior during e-evaluation. The proposal doesn't need additional components and considered the continuous authentication using the random generation of a variable number of questions built from the collected information. The concerned student must respond to these questions which are generated periodically throughout the e-evaluation for guaranteeing his continuous authentication. Simulation experiments are conducted to validate our approach in which the obtained results show that student authentication is effectively guaranteed at a low cost whatever the student number and whatever the e-evaluation content.