Towards a framework for building automatic recommendations of answers in MOOCs' Discussion Forums

Abstract

In this paper, we propose a framework for building automatic recommendations of answers in MOOCs' Discussion Forums. The general aim of this research is to contribute to addressing the MOOCs' dropout rate issue due to the lack of interactivity and scaffolding for learners especially in discussion forums. To this end, the proposed system provides learners with automatic adequate answers to the inquiries and questions they posted in discussion forums. The proposed answering approach is accomplished in two main phases: The Answering System identifies firstly a learner's question among a panoply of exchanged messages within discussion forums. Then, it provides to the learner suitable suggested answers based on similar existing posts, which would help the learner getting immediate preliminary answers without the need to wait for a formal feedback from the instructor. In a second hand, the system builds and recommends an appropriate answer using MOOCs' videos, based on automated crawling and indexing methods and applying web mining and recommendation strategies.