Automated Acquisition of Domain Model for Adaptive Collaborative Web-based Learning

Marián Šimko

Doctoral thesis project supervised by prof. Mária Bieliková

Motivation and Goals

Massive spread of Web 2.0 technologies altered the face of web-based learning. They enabled to realize the "Read/Write Web" vision, where a user is no longer a passive consumer of information, and reflect it into learning. Web 2.0 principles and technologies improved user experience resulting into more interaction, communication, collaboration and extended competences during learning. The paradigm shift affects adaptive learning, where learning experience differs among an individual learner. Delivered learning material is tailored according to user goals, needs and characteristics resulting into more efficient learning, where a learner is able to learn efficiently. An adaptive educational system necessitates semantic descriptions of learning resources to enable machines to “understand” the content, to perform at least a basic reasoning and to provide an adaptive experience. This need is particularly emphasized when considering adaptive web-based learning 2.0, i.e., adaptive learning in dynamic environments with user-generated content being created on daily basis.

In the thesis we dealt with the problem of automated creation of domain model for adaptive web-based collaborative educational systems. The analysis of state-of-the-art approaches to domain modeling for adaptive web-based educational systems revealed the following open problems concerned with domain model and approaches to its creation:

  • no clear separation of conceptualization and content. There is no clear distinction between domain conceptualization and the content of resources being presented in the state-of-the-art adaptive web-based systems.
  • limited support for "2.0" aspects of learning. The increasing importance of a user as a one of central concepts of Web 2.0 paradigm, where he is not only the consumer of information but also a creator of content, is still not sufficiently reflected into the domain model of adaptive web-based system.
  • limited results of methods for domain model composition automation. There is a small number of approaches to automatic creation of domain model for adaptive web-based systems not yielding satisfactory results.

We aim to address these problems by devising methods for supporting automated domain model creation for adaptive collaborative educational web-based systems. In particular, thesis goals are:

  • to design and evaluate a domain model addressing open problems in the state-of-the-art, particularly focusing on separation of concerns and explicit support for modeling annotating and collaborating over the learning content while supporting automated creation of domain model parts.
  • to design and evaluate a method for automated domain model creation by leveraging selected methods and techniques of data mining as well as specifics of educational content and learning in general while considering heterogeneous sources of information.


The core contributions of our work addressing the rigidity of domain modeling with respect to social collaborative learning and the lack of methods supporting automated or automatic domain model creation are:

  • proposal of lightweight domain model for adaptive collaborative web-based learning. We proposed and evaluated new generation domain model for adaptive collaborative educational systems which clearly separates between domain conceptualization and content, explicitly supports collaborative interactive learning, and is proposed with regard to automated creation and enrichment facilitated by leveraging user-generated content, while preserving accurate adaptation.
  • method for automated domain model acquisition. We proposed and evaluated method for automated domain model acquisition that is based on statistical and linguistic processing of underlying resource content. The method consists of several steps that cover the whole knowledge acquisition process ranging from resource preprocessing to final domain model fine-tuning performed by a teacher. The particular method steps represent partial contributions for the field of adaptive educational systems authoring.

In addition, an important contribution as a realization and practical result of our research is:

  • adaptive learning framework ALEF and its instance already used in several courses to support learning. The author of this work actively participated in and co-led the development of adaptive learning framework ALEF, which is a practical result of a research conducted in the area of adaptive web-based learning presented in the thesis. ALEF is an ultimate education supporting software solution merging adaptive learning, collaborative learning and interactivity built on the concepts of Web 2.0. The proposed lightweight domain model is one of the three core pillars of the framework (together with the extensible personalization and collaboration support).


The proposed lightweight domain model constitutes the first step towards genuine adaptive social learning, where the learning content is not created solely by the teachers, but is generated by its users as well. In order to deliver adapted presentation or navigation of user-generated content and make it meaningful for the learning process, novel methods for personalization and group adaptation, reaching beyond traditional learning content or navigation adaptation, need to be devised.

In our method for automated acquisition of relevant domain terms and relationships between them we particularly focused on two types of relationships. However, also other relationship types can be advantageous for the domain model and can result in enhanced adaptation or other intelligent functionality. Additional relationship types also represent a challenge of seeking methods for acquiring relationships of such types (semi-)automatically as well.

An interesting research direction resides in interconnection of the lightweight domain model with domain models for open information spaces in order to allow for unified domain and user modeling in the Web with a perspective of shifting owards the Web and foster more efficient learning on the Web, hence supporting lifelong learning not only in relation to the temporal aspects of learning, but also in the context of e-learning 2.0 and even beyond, by providing more competences for a learner and continuously enhancing the learning process.

The thesis extended abstract is available in the Bulletin of the ACM Slovakia.

Selected publications

Šimko, M., Bieliková, M.
Discovering Hierarchical Relationships in Educational Content. In LNCS 7558, Proc. of Int. Conf. on Web-based Learning, ICWL 2012. Springer. To appear. (2012)
Šimko, M., Barla, M., Mihál, V., Unčík, M., Bieliková, M.
Supporting Collaborative Web-Based Education via Annotations. In Proc. of World Conference on Educational Multimedia, Hypermedia & Telecommunications, ED-MEDIA 2011. AACE, pp. 2576–2585 (2011)
Móro, R., Srba, I., Unčík, M., Bieliková, M., Šimko, M.
Towards Collaborative Metadata Enrichment for Adaptive Web-Based Learning. In Proc. of Int. Workshop on Computational Social Networks, Web Intelligence 2011. IEEE, pp. 106–109 (2011)
Lučanský, M., Šimko, M., Bieliková, M.
Enhancing Automatic Term Recognition Algorithms with HTML Tags Processing. In Proc. of Int. Conf. on Computer Systems and Technologies, ComSysTech’11. ACM, pp. 173–178 (2011)
Šimko, M., Barla, M., Bieliková, M.
ALEF: A Framework for Adaptive Web-based Learning 2.0. In Reynolds, N., Turcsányi-Szabó, M. (Eds.): KCKS 2010, IFIP Advances in Information and Communication Technology, Vol. 324. Held as Part of World Computer Congress 2010. Springer, pp. 367–378 (2010)
Barla, M. Bieliková, M., Bou-Ezzedine, A., Kramár, T., Šimko, M, Vozár, O.
On the Impact of Adaptive Test Question Selection for Learning Efficiency. Computers & Education, Vol. 55, Issue 2, pp. 846–857 (2010)
Šimko, M. Bieliková, M.
User Modeling Based on Emergent Domain Semantics. In LNCS 6075, Proc. of User Modeling, Adaptation and Personalization, UMAP 2010, Doctoral Consortium. Springer, pp. 411–414 (2010)
Šimko, M. Bieliková, M.
Automated Educational Course Metadata Generation Based on Semantics Discovery. In LNCS 5794, Proc. of European Conf. on Technology Enhanced Learning: Learning in the Synergy of Multiple Disciplines, EC TEL 2009. Springer, pp. 99–105 (2009)

to Homepage to Teaching to the Top

Last updated:
Mária Bieliková bielik [zavináč] fiit-dot-stuba-dot-sk
Design © 2oo1 KoXo