Content uploaded by Maurice Hendrix
Author content
All content in this area was uploaded by Maurice Hendrix
Content may be subject to copyright.
KONG, S.C., et al. (Eds.), ICCE2009; ©2009 Asia-Pacific Society for Computers in Education.
Design and Development of an Authoring Tool
for Pedagogical Relationship Types between
Concepts
Dietrich Albert
a
, Alexander Nussbaumer
a
, Christina M. Steiner
a
,
Maurice Hendrix
b
, Alexandra Cristea
b
a
Department of Psychology, University of Graz, Austria
b
Department of Computer Science, University of Warwick, United Kingdom
dietrich.albert@uni-graz.at
Abstract: In this paper the design and development approach of an authoring tool for
defining pedagogical relationship types is described. Pedagogical relationship types are
used to define adaptation strategies which can be used by an adaptation engine to create an
adaptive course based on user model information. Based on domain models consisting of
concepts and relationships between concepts, pedagogical relationships can be used to
connect concepts of domain models in a pedagogical meaningful way. Each type of these
relationships contains adaptation code which formally specifies the meaning of relationships
on a technical and a pedagogical level. To support authors in creating these relationship
types a tool is being developed which easily allows the authoring process.
Keywords: adaptation strategies, concept relationship, authoring tool, adaptive engine
Introduction
After more than a decade of research activities in the field of educational adaptive systems
(for example [2]) an effort in the context of the EC-funded project GRAPPLE is currently
undertaken to develop a generic adaptation approach which can be used in popular Learning
Management Systems (LMS) [7]. For this reason the GRAPPLE Adaptive Learning
Environment (GALE) is being developed and integrated with the major LMSs (including
Moodle and Sakai) using a service-oriented architecture approach. While typical LMSs
often have functionalities for managing learning material and courses, they often have only
little or no features for providing learning content in an adaptive way. At this point GALE
comes into play, since it is designed to provide adaptation functionalities, such adaptive
guidance through link generation and annotation, or adaptive page content to automatically
compensate missing prerequisite knowledge.
In order to fulfil advanced adaptation tasks, GALE is equipped with a domain model
and user model infrastructure. For the sake of integration these models are connected with
information available in the LMS. While actual resources and general information on users
are available in the LMS, the GALE manages conceptual knowledge on subject domains
(domain models) and information on learning progress of users regarding domain
knowledge (user models). Obviously conceptual domain knowledge and pedagogical
strategies used for adaptation behaviour of GALE have to be created by a content author.
After describing related work, this paper presents the GRAPPLE approach how
pedagogical strategies can be created upon conceptual knowledge using pedagogical
KONG, S.C., et al. (Eds.), ICCE2009; ©2009 Asia-Pacific Society for Computers in Education.
relationship types (also called concept relationship types (CRT)). Then the formal definition
of a CRT and the implementation of the authoring tool is presented. A conclusion and an
outlook on further work are given in the last section.
1. Related Work
1.1 Concept relationship types in AHA! and their use in the Graph Author tool
When the AHA! system [5] was initially designed, the actions of the user resulted in user
model updates, which in turn resulted in adaptation through a set of event-condition-action
rules (ECA rules) that had to be created by the author of the adaptive application (or course).
This was a laborious process because it required a lot of repetitive work to author many
instances of identical or very similar user model and adaptation behaviour. At the same
time, it was also difficult for non-technical authors to create ECA rules in order to define the
adaptation for their course material.
AHA! Version 2 (and later 3) introduced concept relationships (CRs) and concept
relationship types (CRTs) [4]. A CRT represents a type of relationship between a set of
generic concepts (or placeholders or variables for concepts). A CR is an actual relationship
between a set of specific concepts (or real concepts from the adaptive application). This is
how CRs and CRTs are defined in the AHAM reference model [6]. In AHA! a CR or CRT
is limited to being a unary or binary relationship (type).
1.2 The Domain Map Editor for Defining Skills and Skill Structures
The Domain Map Editor is a tool for structuring knowledge domains and for defining
curricula. It has been developed in the FP6 research project iClass (http://www.iclass.info/)
in conjunction with other skill-based learning tools (for planning, competence assessment,
self-evaluation, learner knowledge presentation) which make use of the structured data [1].
The base elements of these tools are the skill and the prerequisite relation between skills.
Prerequisite relations between skills are the most important information used for the
adaptation process. A prerequisite relation between skills expresses the psychological
dependence between skills. If a person has available a specific skill, then - due to
psychological reasons - this person also has available all prerequisite skills. Therefore a
learner should acquire skills in a sequence which relies on the fact that for each skill all
prerequisite skills should be taught before. In this way a meaningful sequence of learning
objects can be created automatically. In the light of CRTs there is only one relationship type,
which is the prerequisite relationship between skills. This type cannot be altered and it is not
possible to add new relationship types.
1.3 My Online Teacher
My Online Teacher (MOT) (http://www.dcs.warwick.ac.uk/~acristea/mot.html) is an
adaptive hypermedia authoring system for on-line adaptive course production, which can
provide specifications of adaptation for various types of user-model and presentation-model
related adaptations. MOT is based on the Layers of Adaptive Granularity (LAG) model [3]
which has been introduced as a model for the adaptive behavior within adaptive hypermedia
(AH). LAG is destined for authors of AH entering this process at the different levels of
difficulty and flexibility: direct adaptation rules, adaptation language and adaptation
KONG, S.C., et al. (Eds.), ICCE2009; ©2009 Asia-Pacific Society for Computers in Education.
strategies. LAG allows modeling reusable adaptation at different levels: beginner authors
can reuse entire strategies, by just reading their description. The collection of adaptation
strategies is not unlike how CRTs are supposed to be stored, based on their name, their
description, and their code. The other similarity is the reusability factor and their generality
(can be applied for any domain models). For authors with more experience, direct
programming in the adaptation language is possible, thus allowing both altering of extant
strategies, as well as the creation of new ones. Again, this is similar to the CRT
editing/modifying/deleting.
2. Pedagogical Relationships between Concepts
2.1 Using pedagogical relationships to define an adaptation strategy upon a domain model
The domain model (DM) defines and represents a knowledge domain. Basically, the domain
model consists of a set of concepts and relations between the concepts (concept map, see
Figure 1). The DM does not represent actual content resources (hyper documents or learning
objects), but is rather the underlying representation of the domain to which content
resources can be related. This means that the DM represents the domain on a conceptual
level, and is defined completely independent from actual content resources. The DM can be
created by a teacher/course author familiar with the subject domain, and is visualised in
form of a subject matter graph (conceptual graph), where the nodes represent the concepts
and the edges represent the relations between concepts.
Basically, three different categories of relationships (or relationship types) can be
distinguished and are involved in the context of the authoring tools: hierarchical relations,
semantic relations, and pedagogical relations. While hierarchical relations and semantic
relations among domain concepts are dealt with in the DM, the pedagogical relations
actually refer to the relationship types that are captured by the CRT. These pedagogical
relationships are not used to describe the domain on a semantic level, but constitute types of
relations that are used for defining adaptation rules and that are subsequently used for
realising instructional strategies. By using concepts from the DM and connecting them with
pedagogical relationships the Concept Adaptation Model (CAM) is created which is used by
the adaptation engine for the adaptive course (see Figure 1).
Special cases are, however, semantic relations or hierarchical relations of the domain
model that are used for pedagogical purposes. Such relations need to be associated a certain
adaptation behaviour, which is possible also via the CRT tool. An example is, for instance,
the usage of the hierarchical is-a relation between concepts to present them in breadth-first
or depth-first manner, as required by the holistic and sequential learning styles, respectively.
KONG, S.C., et al. (Eds.), ICCE2009; ©2009 Asia-Pacific Society for Computers in Education.
Figure 1: The graph on the left side depicts a domain model and the graph on the right side shows how
concepts from the domain model are connected with the pedagogical "prerequisite" relationship type.
2.2 Concept Relationship Types
The pedagogical relationships described above are instances of concept relationship types.
In contrast to the pedagogical relationships which are used between concrete concepts, the
concept relationship types (CRT) are the formal specifications of pedagogical relationships.
They define the structure how (and how many) concepts can be connected and most
importantly they also define the meaning of a pedagogical relationship.
In order to define the meaning each CRT has to be assigned with a piece of adaptation code
which can be interpreted by the adaptation engine. In this way the adaptation behaviour can
be specified for each CRT which defines the behaviour of the adaptation engine. The
language for this piece of code is the GRAPPLE adaptation language (GAL) which is
currently developed. For example, the GAL code for prerequisite relationship type will be
something like "if (user has visited concept A) then (concept B is suitable)". As seen in this
example, the user model variables can be accessed using the adaptation language.
Information of a user, such as visiting state of a concept or knowledge level of a concept can
be used to formally define the adaptation behaviour. In this way the pedagogical meaning
can be formally expressed.
3. Implementation of an Authoring Tool for Concept Relationship Types
The CRT authoring tool consists of a Web-based tool which provides the possibility to
create and define CRTs and a Web Service where the created CRTs can be stored and
retrieved by other tools. Authors can input the information described above in a graphical
way and save them to the Web Service. For the reason of interoperability the CRT data is
expressed in an XML-based format and the certain CRTs are saved as XML files on the
server side into a database behind the Web Service. The graphical tool is implemented in
Adobe Flex 3 technology (http://www.adobe.com/de/products/flex/) and the Web Service is
realised with Apache Axis2 framework (http://ws.apache.org/axis2/).
The information needed to specify a CRT includes:
• general information of CRTs, such as name and description,
• information how an instance of a CRT should be visually represented in the CAM
• the GAL code which formally defines the adaptation behaviour
• the user variables which the GAL code is accessing
• the properties of the source and target socket which the CRT connects
KONG, S.C., et al. (Eds.), ICCE2009; ©2009 Asia-Pacific Society for Computers in Education.
• constraints, such as the information if sequences of CRT instances may form a loop
(which would not make any sense for prerequisite relationships) or if concepts with
specific attributes are excluded
• some technical information such as creation time or author
The most challenging part of the tool development is to provide as much as possible support
to the author who wants to create and modify adaptation code. Programming adaptation
code directly in GRAPPLE adaptation language excludes most content authors who do not
have programming skills available from creating their own CRTs. Therefore, this authoring
tool shall provide graphical techniques, which allows an inexperienced author to create new
CRTs easily.
4. Conclusion and Outlook
In this paper the design and development approach of an authoring tool for pedagogical
relationship types has been described. Pedagogical relationship types are used to define
adaptation strategies which can be used the adaptation engine to create an adaptive course
based on user model information.
Further development will focus on providing visual support for authoring the
adaptation behaviour which is expressed in programming code. This should enable beginner
authors to create new CRTs or to modify existing ones. Evaluation of this tool will be done
in the context of the GRAPPLE project internal evaluation of the authoring tools.
Acknowledgements
This paper and the work presented in this paper is part of the ongoing research and
development in the EU FP7 project GRAPPLE (Project Reference: 215434) and could not
be realized without the close collaboration between all 15 GRAPPLE partners, not listed as
authors, but nonetheless contributing to the ideas described here
References
[1] Albert, D., Nussbaumer, A., & Steiner, C. (2008). Using Visual Guidance and Feedback Based on
Competence Structures for Personalising E-Learning Experience. Proceedings of the 16th International
Conference on Computers in Education (ICCE 2008), 27-31 October 2008, Taipei, Taiwan.
[2] Brusilovsky, P. (1996) Methods and techniques of adaptive hypermedia. User Modeling and
User-Adapted Interaction, 6 (2-3), pp. 87-129
[3] Cristea, A.I., & Calvi, L. The three Layers of Adaptation Granularity. International Conference on User
Modelling (UM 2003), Pittsburgh, US, Springer, 2003
[4] De Bra, P., Aerts, A., Rousseau, B. (2002).Concept Relationship Types for AHA! 2.0. In M. Driscoll, &
T. C. Reeves (Eds.), Proceedings of World Conference on E-Learning, E-Learn 2002, October 15-19,
2002. Montreal, Canada, AACE, 1386-1389.
[5] De Bra, P. & Calvi, L. (1998). AHA! An open Adaptive Hypermedia Architecture. The New Review of
Hypermedia and Multimedia, vol. 4, pp. 115-139, Taylor Graham Publishers.
[6] De Bra, P., Houben, G.J., Wu, H. (1999).AHAM: A Dexter-based Reference Model for Adaptive
Hypermedia. Proceedings of the ACM Conference on Hypertext and Hypermedia, 147-156. Germany:
Darmstadt.
[7] De Bra, P., Pechenizkiy, M., van der Sluijs, K., & Smits, D (2008). GRAPPLE: Integrating Adaptive
Learning into Learning Management Systems. Proceedings of the World Conference on Educational
Multimedia, Hypermedia & Telecommunications (ED-MEDIA 2007), July 2008, Vienna, Austria.
