Here's the starting point: "Despite the advantages MOOCs offer, MOOCs have high dropout rates, and literature suggests this may be tied to participant motivation." The idea here (21 page PDF) is to use topic modeling to identify participant motivations for taking a MOOC, and in turn use this to predict whether they'll drop out. Three different methods were used, with varying results. The discussion mostly revolves around a comparison of the three approaches, and I'm sort of left wondering whether they ever answered the question they posed at the start, and ultimately "the goal of this study was to investigate how topic modeling can be used for analyzing qualitative data," to which the answer is, 'maybe'.
Today: 0 Total: 17 [Direct link] [Share]
