Data-driven analyses of electronic text books

Research output: Contributions to collected editions/worksContributions to collected editions/anthologiesResearch

Authors

We present data-driven log file analyses of an electronic text book for history called the mBook to support teachers in preparing lessons for their students. We represent user sessions as contextualised Markov processes of user sessions and propose a probabilistic clustering using expectation maximisation to detect groups of similar (i) sessions and (ii) users. We compare our approach to a standard K-means clustering and report on findings that may have a direct impact on preparing and revising lessons.
Original languageEnglish
Title of host publicationSolving large scale learning tasks : Challenges and algorithms : essays dedicated to Katharina Morik on the occasion of her 60th birthday
EditorsStefan Michaelis, Nico Piatkowski, Marco Stolpe
Number of pages15
Place of PublicationCham
PublisherSpringer International Publishing AG
Publication date2016
Pages362-376
ISBN (print)978-3-319-41705-9
ISBN (electronic)978-3-319-41706-6
DOIs
Publication statusPublished - 2016