Abstract | ||
---|---|---|
As online courses such as MOOCs become increasingly popular, there has been a dramatic increase for the demand for methods to facilitate this type of organisation. While resources for new courses are often freely available, they are generally not suitably organised into easily manageable units. In this paper, we investigate how state-of-the-art topic segmentation models can be utilised to automatically transform unstructured text into coherent sections, which are suitable for MOOCs content browsing. The suitability of this method with regards to course organisation is confirmed through experiments with a lecture corpus, configured explicitly according to MOOCs settings. Experimental results demonstrate the reliability and scalability of this approach over various academic disciplines. The findings also show that the topic segmentation model which used discourse cues displayed the best results overall. |
Year | Venue | Field |
---|---|---|
2015 | EDM | Data science,Segmentation,Computer science,Discipline,Artificial intelligence,Machine learning,Scalability |
DocType | Citations | PageRank |
Conference | 1 | 0.35 |
References | Authors | |
15 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ghada AlHarbi | 1 | 8 | 1.90 |
Thomas Hain | 2 | 171 | 28.29 |