An Intelligent Examination Monitoring Tool for Online Student Evaluation

Abstract

The global reach of online education has increased due to a pandemic or other unique circumstances. As online education got more popular, it became crucial to ensure the quality of evaluation. This study's goal is to find a solution to the issue of monitoring during online exams. We have used behavioural biometrics through students' interaction with an Intelligent Examination Monitoring Tool (IEMT), which was developed, even though many studies concentrate on using video analysis. The test-taking prototype uses mouse, touch, and keyboard interfaces to administer multiple-choice questions with a variety of information and events. Students who used additional sources to answer questions were later discovered during an online interview. We built a prediction model that can determine if a student is answering on his own or using any other sources using the events through input interaction when these students are sorted. The Machine Learning (ML) techniques Decision Tree, Random Forest, K-Nearest Neighbour, and Naive Bayes were used to generate a few models. After evaluating the performance of the models, we find that random forest performs best, with an accuracy of about 91 percent.