It Is challenging for the teacher to identify which students are having difficulties in online learning. Indicators like the number of forum posts or the students' reflection on feedback can be extracted from the activity logs in the Learning Management Systems. This thesis describes using supervised learning algorithms to automatically classify students In a course, taking these indicators as input. Multiple neural network architectures have been developed and compared in this thesis. Their accuracy using data from new courses and data from new sites ls In the 71.30-83.09% range. The developed framework has been integrated with the Moodie LMS.
|Award date||20 Aug 2019|
|Publication status||Unpublished - 2019|