dataTEL - Datasets for Technology Enhanced Learning. White paper

The dataTEL white paper develop during the dataTEL workshop at the ARV2011. The workshop was motivated by the issue that very less educational datasets are publicly available in TEL, so that the outcomes of different TEL adaptive applications and recommender systems that support personalised learning are hardly comparable. In other domains like in e-commerce it is a common practise to use different datasets as benchmarks to evaluate recommender systems algorithms to make the results comparable (MovieLens, Book-Crossing, EachMovie dataset). So far, no universally valid knowledge exists in TEL on algorithm that can be successfully applied in a certain learning setting to personalise learning. Having a collection of datasets could be a first major step towards a theory of personalisation with in TEL that can be based on empirical experiments with verifiable and valid results. Therefore, the main objective of the dataTEL workshop was to explore suitable datasets for TEL with a specific focus on recommender and adaptive information systems that can take advantage of these datasets. In this context, new challenges emerge like unclear legal protection rights and privacy issues, suitable policies and formats to share data, required preprocessing procedures and rules to create sharable datasets, common evaluation criteria for recommender systems in TEL and how a dataset driven future in TEL could look like.