There's a lot of creativity at work in this article (21 page PDF). Essentially, the objective is to create a mechanism of classifying MOOCs in order to be able to say something about their quality, but there are no good frameworks to speak of. So the authors took a BERT classifier, trained it on Bloom's Taxonomy, and turned it loose on some MOOCs. The authors dive deep into how this is done, but the upshot is that they come out with an analysis of 2394 learning objectives based on Bloom's, evaluating various applications of BERT to perform the work, resulting in "a model for the automatic classification of MOOCs according to their pedagogical approaches."
Today: 7 Total: 83 [Direct link] [Share]
