Time-dependent recommender systems for the prediction of appropriate learning objects
Published | June 2018 |
Type of work | Doctoral Thesis |
Periodical | Pages 1-301 |
Institution | Technische Universität Berlin |
Advisor | Hauswirth, Manfred |
Country | Germany, Europe |
ABSTRACT
This dissertation deals with adaptive learning technologies which aim to optimize Technology Enhanced Learning (TEL) offerings to fit the individual learner’s needs. Thereby, Recommender Systems play a key role in supporting the user’s decision process for items of interest. This works very well for e-commerce and Video on Demand services. However, it is found to be the case that these traditional Recommender Systems cannot be directly transferred to TEL as the recommendation of course items follows a particular educational paradigm. The special conditions of this paradigm are first investigated and then taken into account for the realization of new algorithms.In order to allow a broad interoperability of a Recommender System with other technical components, a set of open standards and specifications results in a reference architecture for such an adaptive learning environment. Based on the realized architecture, activity data have been collected from students using course materials available online – the courses themselves comprising face-to-face lectures backed up by digital representations of the presented contents, blended learning settings as well as online-only courses. The courses provided access to the course materials via a novel Learning Companion Application. This app also presents learning recommendations to make the content selection more efficient and effective.Thereby, this work indicates that an educational Recommender System should not be evaluated using standard evaluation frameworks that utilize, for instance, a classical n-fold cross-validation. For this reason, a time-dependent evaluation framework is defined to investigate the precision of the Top-N learning recommendations at various points in time. Moreover, a new measure is introduced to determine the Mean Absolute Timeliness Deviation between an item recommendation and the time when it is actually accessed by the user. Subsequently, four major techniques for Recommender Systems are realized and applied to the collected data, evaluated with the time-dependent evaluation framework and successively optimized. As a reference implementation, a traditional Collaborative Filtering algorithm is developed and extended to incorporate time information. The results are then compared to the results of other time sensitive algorithms: an Item-based Collaborative Filtering approach which has previously been applied to TEL and a new learning path generator which incorporates a set of contextual information. Finally, a novel time-weighted Knowledge-based Filtering algorithm is presented and exhaustively analyzed. The evaluation results indicate that time-dependent filtering which incorporates multi-contextual activity data can produce the most precise recommendations.
ABSTRACT
Die vorliegende Dissertation beschäftigt sich mit adaptiven Lerntechnologien, die sich an die individuellen Bedürfnisse der Lernenden anpassen. Dabei spielen vor allem Empfehlungssysteme eine Schlüsselrolle, da sie den Entscheidungsprozess der Benutzer unterstützen. Das funktioniert sehr gut für E-Commerce und Video on Demand-Dienste. Allerdings können diese Mechanismen nicht einfach für den Bereich des Technologie-gestützten Lernens übertragen werden, da die Empfehlungen von Kursinhalten einem sehr speziellen Paradigma folgen. Die Eigenschaften dieses Paradigmas werden in der Dissertation erst analysiert und anschließend als Basis für neue Algorithmen berücksichtigt. Um eine breite Interoperabilität des Empfehlungssystems mit anderen technischen Komponenten zu gewährleisten, wurden offene Standards und Spezifikationen umgesetzt, mit deren Hilfe eine Referenzarchitektur für adaptive Lernumgebungen umgesetzt wurde. Basierend darauf wurden Aktivitätsdaten in Echtwelt-Kursen gesammelt – von Präsenzunterricht, welcher durch digitales Vorlesungsmaterial unterstützt wurde, über Blended Learning-Umgebungen bis hin zu ausschließlichen Online-Kursen. Alle Kursteilnehmer hatten Zugriff auf die Kursmaterialien über die Lernbegleiter-App. Der Entscheidungsprozess der Lernenden wurde durch ein Lernempfehlungssystem unterstützt. Dabei hat sich herausgestellt, dass herkömmliche Evaluationstechniken, wie die n-Fold Cross- Validation, nicht für die Evaluation von Lernempfehlungssystemen geeignet sind. Deshalb wurde ein zeitabhängiges Evaluations-Framework definiert, mit dem die Präzision von Top-N-Lernempfehlungen zu verschiedenen Zeitpunkten analysiert werden kann. Zusätzlich wurde eine neuartige Messgröße eingeführt, die „Mean Absolute Timeliness Deviation”, die den zeitlichen Abstand zwischen Empfehlungen und dem späteren Abruf der Inhalte durch den Benutzer misst. Darauf basierend konnten vier Haupttechniken für Empfehlungssysteme realisiert und auf die gewonnenen Datensätze angewandt werden. Dann wurden diese mit dem definierten Evaluations- Framework ausgewertet und sukzessive optimiert. Als Referenzimplementierung diente ein traditioneller.Published at | Berlin, Germany |
Language | German |
Refereed | Does not apply |
Rights | Attribution 4.0 International (CC BY 4.0) |
DOI | 10.14279/depositonce-7119 |
URL | https://depositonce.tu-berlin.de//handle/11303/7957 10.14279/depositonce-7119 |
Export options | BibTex · EndNote · Tagged XML · Google Scholar |
Viewed by 105 distinct readers
CLOUD COMMUNITY REVIEWS
The evaluations below represent the judgements of our readers and do not necessarily reflect the opinions of the Cloud editors.










Click a star to be the first to rate this document
▶ POST A COMMENT
SIMILAR RECORDS
OER’s in Sub-Saharan Africa – An appropriate response to the challenge of education for all? The TESSA experience in Zambia.
Dart, Gareth
The initiatives of Education For All (EFA) and the Millennium Development Goals (MDGs) have helped put the debate surrounding basic education firmly in the educational, developmental and economic spotlight. They have ...
Match: appropriate
The future of MOOCs: Adaptive learning or business model?
Daniel, John; Vázquez Cano, Esteban; Gisbert, Mercè
Currently, many MOOCs are designed as a collection of videos with a forum using some traditional distance learning models, but they do not promote adaptive and personalized learning. These features, together with the ...
Match: adaptive learning
Open educational resources: Global report 2017
Three documents have been developed in the framework of the 2nd World OER Congress. The two documents: Open Educational Resources: From Commitment to Action and the Open Educational Resources: Global Report 2017 provide ...
Match: europe
ICDE media release- December 10, 2014
International Council for Open and Distance Education (ICDE)
A press release relating the results from discussions at the ICDE - UNESCO Policy Forum in Bali, November 2014.
Match: technology enhanced learning
MOOCS Massive Open Online Courses
Gaebel, Michael
Match: ; europe
Massive Open and Online Courses (MOOC) in an international perspective: New global agenda for innovation in higher education
Titlestad, Gard
This report is written based on a note provided by ICDE’s Secretary General Gard Titlestad to the Norwegian Government. It has been slightly modified to adapt it for an international audience and a working ...
Match: europe
Are we having fun yet? Institutional resistance and the introduction of play and experimentation into learning innovation through social media
Bryant, Peter; Coombs, Antony; Pazio, Monika
Recognising and responding to behaviours and patterns of resistance is critical to the successful implementation of technology-enhanced learning strategies at higher education institutions. At institutional, academic ...
Match: technology enhanced learning
Porto declaration on European MOOCs
During the HOME conference Mapping the European MOOC Territory held on 27 November 2014 in Porto, participants discussed and agreed the text of a declaration about MOOCs and the opportunities they present in ...
Match: ; europe
Design framework for an adaptive MOOC enhanced by blended learning: Supplementary training and personalized learning for teacher professional development
Gynther, Karsten
The research project has developed a design framework for an adaptive MOOC that complements the MOOC
format with blended learning. The design framework consists of a design model and a series of learning design ...
Match: adaptive learning
JiFUNzeni: A blended learning approach for sustainable teachers professional development
Onguko, Brown Bully
JiFUNzeni blended learning approach is a sustainable approach to provision of professional development (PD) for those in challenging educational contexts. JiFUNzeni approach emphasizes training regional experts to ...