Search

Past research on analysing end-of-term student feedback tend to result in only high-level course improvement suggestions, and some recent research even argued that student feedback is a poor indicator of teaching effectiveness and student learning. Our intelligent Q&A platform with machine learning prediction and engagement features allow students to ask self-directed questions and provide answers in an out-of-class informal setting. By analysing such high quality and truthful posts which represent the students’ queries and knowledge about the course content, we can better identify the exact course topics which the students face learning challenges. We have implemented our Q&A platform for an undergraduate spreadsheets modelling course, and analysed 1025 meaningful posts to identify the hot areas represented as topic tags, map the identified hot tags progression over time, to direct instructors towards targeted improvement actions. Our proposed approach can be applied to other courses where students’ self-directed Q&A can be implemented.

We designed, developed and implemented an integrated Telegram and Web-based Forum capable of automatic thoughtfulness assessment of questions and answers, using text mining and Natural Language Processing (NLP) methodologies, for collaborative learning. We trained and applied the Random Forest algorithm to provide instant thoughtfulness score prediction for the new posts contributed by the students, and prompted the students to improve on their posts, thereby invoking deeper learning. In addition, the platform is designed to ensure that students remain engaged on the platform. Our initial implementation for a university course on spreadsheets modeling using Excel, shows some promising results which strengthened our belief that through the process of asking good questions, and providing good answers, students' learning can be improved beyond the traditional classroom, and beyond instructor-student relationship.