Hendrik Thüs

 

Learning Context: A Privacy Preserving Platform for Learning Data Collection and Analytics

Nowadays, learning does not only happen in a classroom anymore. Due to the vastly growing number of smartphone / tablet users, learning can take place everywhere and every time. Such a scenario creates many new opportunities to improve learning by making use of the various kinds of context information that can be gathered by modern (mobile) devices.

The Learning-Context project aggregates different sets of personalized and learner-related data in a model which can be used to give valuable feedback or recommendations or to allow self-reflection on the daily tasks or learning activities. Such a model, called the LCDM, short for Learning Context Data Model, should reflect a complete picture of the learner, starting with context information and extending it with interests, knowledge, goals, background, and individual traits coming from various sources.

The first main challenge in this dissertation is to design the LCDM which is able to hold the previously mentioned set of learner-related information while avoiding unnecessary entities and relations and preserving the semantics of the data. Since the LCDM should evolve over time, it should be flexible enough to adapt to new requirements. This flexibility should be shown by extending the solely context-driven data model with interests of the learners. Those interest items should be able to evolve over time and therefore represent the history and the current status of a learner’s interests.

The second main challenge is to design and implement an API which is capable of giving third-party applications access to the LCDM. In order to fulfill the requirements of an Open Learner Model, each learner should be able to have full control of his/her own data. For one thing, this requires adding, requesting, editing and deleting functionalities, for another thing, the learner should be able to decide who or which application has revocable (pseudonymized) access to his/her data. The design and the implementation of the API should address the question how the learners can be convinced of the advantages of the Learning Context project.