Educational Data Mining: Machine Learning Techniques for Predicting At-risk of Failure Students

Research on educational data mining has been primarily based on the LMS dataset for evaluating student academic performance. This thesis differs from the body of literature by adding additional datasets that advanced the knowledge of understanding of factors affecting academic performance. The study investigates the application of machine learning techniques based on the internet usage log files and LMS data. Combing internet usage log files and demographic data leads to prompter and more accurate predictions of at-risk students. This study introduces how prediction accuracy can be improved by using a varying range of internet usage data.