Stopout Research Based on Behavior in Massive Open Online Courses

WEN YIN, MING-HUA JIANG, CHANG-LONG ZHOU, LI-LING ZHOU

Abstract


In recent years, the loss of a large number of learners has attracted widespread attention in Massive Open Online Courses (MOOCs). A great deal of research has been modeling analysis and predicted based on the learning behavior. In this paper, according to the modules provided by moocroom, extract some related behavior, select the learners registered after the first three weeks of the data as the data set, using logistic regression modeling analysis, forecast for the next few weeks whether learners stopout, final assessment of the model performance. Experimental results indicate that the model established achieves high prediction accuracy, and the models attained the average AUC is 0.76.

Keywords


MOOC, Stopout, Behavior, Predicted, Logistic regression.


DOI
10.12783/dtcse/cimns2017/17416

Refbacks

  • There are currently no refbacks.