Assessing Online Learners’ Access Patterns and Performance Using Data Mining Techniques

Abstract

Abstract: In an online learning environment where students do not have direct face-to-face interactions with instructors, observing their learning behaviors is quite challenging. As we move from traditional to virtual classrooms, it is important to look into the learning methods and approaches that can be used to have a deeper understanding of student participation in an online learning environment, especially in the Philippine setting where this type of research is relatively very scarce or not investigated at all. This study showed the advances in employing the combination of cluster analysis and data mining in assessing the students' online access patterns and performance. Using the data mining process, this study analyzed the access patterns and performance of undergraduate students in an online course in an open university in the Philippines using the data generated from an open-source LMS called Moodle. Moreover, descriptive and inferential statistics, clustering, and visualization techniques were employed to identify and analyze the students’ behavioral patterns and preferences based on the login frequency, frequency of accessing course materials, number of views and posts in the collaborative learning logs and discussion, announcements, and self-introduction forums, and frequency of submission and completion of assignments, exams, and projects. Results of the study suggest that students are more engaged if given more activities and opportunities for collaboration such as in discussion forums. Results also showed that a big part of students' access to the course site is to access (view and post) the discussion forums and view the introduction forum where they learn something from their classmates. The data mining techniques, particularly the statistics, clustering approach, and visualization can be a useful tool in analyzing the online learners’ access, patterns, and performance.
 Keywords: Data Mining, Learning Management System (LMS), online learning, student performance, clustering