Obviously Educational Data Mining (EDM) has a lot in common with with main stream data mining looking at prediction, clustering and relationship mining. Two items that may not be seen in main stream data mining literture is the distillation of data for human judgement and discovery with model and discovery with models. I am still not sure what this is really about but I believe it is using models such as ontologies to drive the mining strategy.
The paper outlines the key application areas of EDM:
- Improvement of student models - differences in students, how people are learning , who is gaming the system, who is bored or frustrated
- Improve domains knowledge structure
- Used to validate pedagogical support - what pedagogy works and in what circumstances - get a best fit
The paper also discussed public data collections that can now be used to test EDM methods and technologies.
The most popular papers in the area look at relationship mining but in recent years this has not been as popular. The research in vogue in EDM at the moment is prediction - representing 42% of EDM2008 papers.
This paper is a good introduction to the area. It gives good pointers to important papers in the area. It is not particularly thought provoking but I suppose its not meant to be. The next step is to start looking at some of the more important citations make n the paper.