An Open-Source Online Examination System to Meet the Integrity Demands of E-Learning

The rise of online learning platforms and the growing demand for remote education emphasize the importance of online exam-proctoring tools. Online proctoring tools presented in the literature require high internet speed and specialized hardware support, posing accessibility challenges for individuals in developing countries. This study aims to develop a solution that relies on something other than high internet speed and high-end hardware components. The proposed solution extracts data generated from keystroke logs, browser history, and applications opened during the assessment to predict online exam cheating. This data is compared to the words in the test using Term Frequency (TF) and Inverse Document Frequency (IDF) to predict cheating. To evaluate the effectiveness of the proposed solution, an experiment was conducted with sixteen undergraduate Software Engineering students divided into two groups of eight students. The groups were given 20-minute-long software engineering and database exams, each comprising 30 MCQS. These exams were conducted with the proposed proctoring tool and only one group was allowed to cheat. Results indicated that the proposed tool effectively detects cheating during exams. This approach can mitigate the digital divide, particularly for individuals lacking high-speed internet access and costly hardware. Consequently, the study proposes an inclusive solution designed to cater to users from diverse demographic backgrounds.