Personal learning material recommendation system for MOOCs based on the LSTM neural network

Abstract

Massive open online courses (MOOCs; online courses delivered over the Internet) enable distance learning without time and place constraints. MOOCs are popular; however, active participation level among students who take MOOCs is generally lower than that among students who take in -person courses. Students who take MOOCs often lack guidance, and the courses often fail to provide personalized learning materials. Artificial intelligence (AI) has been applied to manage increasing amounts of learning data in learners' online activity records. Driven by the trend in big data, AI technology has drawn increasing attention in various fields. AI -based recommendation systems (RSs) are powerful tools for improving resource acquisition through supply customization, and they can provide personalized learning materials as study guides. In this study, a personalized learning path for MOOCs based on long short-term memory (LSTM) was proposed to meet students' personal needs for learning. According to students' video -watching behaviors, we proposed an MOOC material RS that identifies students with similar learning behaviors through clustering and then uses the clustering results and the learning paths of each group of students to construct an LSTM model to recommend learning paths. The system's learning path recommendations can effectively improve the online participation of learners, and students who received recommendations progressed from the slow -progress group to the medium -progress or fast -progress group. In addition, the learning attitude questionnaire results indicated that the proposed system not only motivated learners to continue learning and achieve high learning capacity but also supported their study planning according to their individual learning needs.