We are happy to announce the next speakers for our itec research seminar

• 12pm-12.10: prof. Dries Benoit (U Ghent – Center for Statistics)
• 12.10-12.50: Wai Kit Tsang (U Ghent)
  
  

Unravelling Drop-out in Fully Asynchronous Subscription-based MOOCs: When is Drop-out Failure or Success?

Predictive models have significantly progressed the understanding of drop-out in Massive Open Online Courses (MOOCs). In the literature it carries a negative connotation as drop-out is considered to stem from low motivation. This study analyses learning behaviour data of a fully asynchronous subscription-based MOOC-platform using hidden Markov models (HMM) and aims to determine the relation between learning motivation and drop-out behaviour. A contribution is made to the literature by establishing a link between goal orientation theory and a latent variable motivation modelled over time. The results surprisingly show that drop-out is not necessarily linked to a low motivation and that students can drop-out despite being very motivated at the end of their learning path. A k-means clustering method was applied to identify two groups of drop-outs, namely the students with very low mastery orientation as opposed to students with a high mastery-orientation that have completed their learning objectives.

• 12.50-13.30: Ekaterina Loginova (U Ghent)

Personalised education with semantic graphs

We develop a semantic approach to data-driven personalisation in education. The goal is to recommend learning content and generate exercises, tailoring both steps to students’ needs and interests. Automating this process lowers adaptive courseware construction costs and allows us to provide wider access to education while maintaining its high quality.

Semantic graphs have not been extensively studied in this context before, despite enjoying success in other fields such as news recommendation. We first aim to develop a semantic graph-based system for adaptive language learning. We will focus on grammar acquisition and context-dependent word usage, using graphs to represent syntactic and semantic data. The next step is to adapt the system for another domain – mathematics. An early prototype of the system is already developed and presented on ICDM.