MOOC student dropout prediction model based on learning behavior features and parameter optimization

Since the advent of massive open online courses (MOOC), it has been the focus of educators and learners around the world, however the high dropout rate of MOOC has had a serious negative impact on its popularity and promotion. How to effectively predict students' dropout status in MOOC for early intervention has become a hot topic in MOOC research. Due to there are huge differences in the learning behaviors, study habits and learning time of different students in MOOC, i.e. the students' learning behavior data containing rich learning information, so it can be used to predict the students' dropout status. In this paper, according to the students' learning behaviour data, a feature extraction method is firstly designed, which can reflect the characteristics of weekly student learning behaviors. Then, the intelligently optimized support vector regression (SVR) model is used as the student dropout prediction (SDP) model. In this SDP model, the three parameters of SVR are not randomly selected but determined by an improved quantum particle swarm optimization (IQPSO) algorithm. Experimental results from both direct observation and statistical analysis on public data indicate that the proposed SDP model can achieve better predictive performance than various benchmark SDP models.

Automated summary

This paper proposes a method to predict students' dropout status using their learning behavior data. By designing a feature extraction method and an intelligently optimized SVR model, this method outperforms other benchmark models in predictive performance.