USING SEMANTIC TECHNOLOGIES AND CASE BASED
REASONING TO SUPPORT COURSE CURRICULUM DESIGN
TASKS
Javier Vaquero, Carlos Toro
VICOMTech Research Centre, Mikeletegi Pasalekua 57, 20009 San Sebastian, Spain
Ricardo Ferrería, Josune Prieto
MIK, Mondragon Unibertsitatea, Onati, Spain
Nieves Alcain, Jesús Rosel
Alecop, Arrasate, Spain
Mar Segura
Department of Computer Science, University of the Basque Country, San Sebastian, Spain
Keywords: CBR, Ontologies, Course Design, Knowledge Based Systems.
Abstract: This paper presents a novel approach to Course Curriculum Design (CCD) where Semantic Technologies
and Case Based Reasoning (CBR) techniques are used to assure (i) a better understanding of the course
being designed and (ii) an efficient use of the available resources. Our work focuses on re-utilization of
previously modelled information (courses, tasks, evaluations, etc.) in order to maximize the efficiency of the
Course Design process while at the same time embedding implicit and experential knowledge of the course
designers. Our approach is presented through an easy to follow architecture that can be adapted to course
curriculum regulations of most European and American models. As a test case, we present an
implementation for a Spanish technician-level telecommunications course, to demonstrate the benefits of
our proposal.
1 INTRODUCTION
A good teaching process should provide the student
with the highest possible quality. It can be argued
that such quality, is strongly related to the best use
of the available resources, the proper design of the
subjects and evaluations, and generally, in the design
of the courses and course components which are part
of the education process (Högskoleverket, 2008).
For the aforementioned reason, the importance of
course/curriculum design is gaining interest for
teachers, education centres and researchers.
It has been reported, that in the present situation
(Rubio Oca, 2006) many curriculum designs do not
met requeriments, due to different factors, e.g. the
evolution of technical tools used when the original
design took place. This fact can lead to problems in
technical adaptation. In many cases the lack of
effective adaptation to the actual situation in
curriculum plans generates a situation where
students are not sufficiently qualified for industry,
implying longer adaptation times when they become
workers.
Typically, a course design starts with the
definition of competences. In other words, the
objectives that must be met at the end of the course
cycle. Diamond (Diamond, 1998) points that
educators need to clearly identify goals prior to any
kind of course assessment. In our case, those goals
are indistinguishable to what we understand as a
384
Vaquero J., Toro C., Ferrería R., Pr ieto J., Alcain N., Rosel J. and Segura M. (2009).
USING SEMANTIC TECHNOLOGIES AND CASE BASED REASONING TO SUPPORT COURSE CURRICULUM DESIGN TASKS.
In Proceedings of the First International Conference on Computer Supported Education, pages 384-387
DOI: 10.5220/0001961203840387
Copyright
c
 SciTePress

competence evaluation. Based on the competences
the course designer builds the content, and later, the
evaluations, producing an output for the students to
follow (a process known as CCD).
Whitin our scope, we have found that CCD
presents several challenges, the following being the
most interesting from a computational perspective:
 Course Curriculum Designers have differing
points of view, which lead to a non-
homogenized, case curriculum.
 The re-use of knowledge and prior user
experiences is not included in the approach.
 Every country has its own course design
legislation. Successful experiences in one
country cannot be easily applied to another.
For these reasons, a computerized system that
aids the competences based CCD is required. We
address this need by presenting a novel approach in
where Semantic techniques are combined with a
Case Based Reasoning (CBR) schema in order to
enhance the precision of the system.
This paper is structured as follows: In chapter 2,
we present an overview of related concepts. In
chapter 3, we introduce our proposed schema using
Semantic technologies and CBR. In chapter 4, we
describe a case study, briefly explaining key points.
Finally in chapter 5, we draw conclusions and
suggest future work.
2 RELATED CONCEPTS
In this chapter, we introduce some concepts relevant
to our work. Our intention is not to provide a
comprehensive description of the topics involved,
but to give a short overview. An interested reader is
invited to review (Fallon and Brown, 2003), (Noy
and McGuiness, 2001), (Aamodt and Plaza, 1994).
2.1 Educational Contents Modelling
e-Learning is defined as “any learning, training or
education that is facilitated by the use of well-known
and proven computer technologies, specifically
networks based on Internet technology” (Fallon and
Brown, 2003).
An important part of the e-Learning process
involves the educational platform. The actual
situation involves propriety design platforms with
their own contents, making interoperability and
interaction between models in use by different
institutions a difficult and considerable task. The re-
use of previous content presents further difficulty
and expense. To rectify this situation, metadata-
based educational standards have been developed.
We believe that important information can be
rendered invalid, or not to be taken into account.
Such information is not directly stored in databases,
and is closely approximated to what we understand
as ‘user experience’.
2.2 Semantic Technologies
In this work, we use ontology modelling for its
inference capabilities and to support our architecture
from a knowledge engineering point of view.
There are many possible definitions to describe
what ontology is. In the Computer Science domain,
the widely accepted definition states that “an
ontology, is the explicit specification of a
conceptualization” (Gruber, 1995), or in other words
an ontology is a description of the concepts and
relationships in a domain of study.
The main characteristic of an ontology based
solution is its capacity to semantically infer newly
derived information. Such information is not
explicitly specified by the user and in order to obtain
it modern inference engines and reasoners, like
Racer or Pellet (Sirin et al., 2007), are used.
2.3 Case Based Reasoning
CBR is a problem solving technique based on two
tenets: (i) the world is regular, so similar problems
have similar solutions, and (ii) types of problems an
agent encounters tend to reoccur (Leake, 1996).
CBR does not use generalized rules as a
knowledge source, but a memory of stored cases
recording specific prior episodes (Leake, 1996).
New solutions are generated by retrieving the most
relevant cases from memory and adapting them to fit
new situations.
We believe that by mixing CBR and Semantic
technologies the strong points of both techniques can
be leveraged to the users advantage.
3 PROPOSED SCHEMA
In this work we propose an open and extensible
architecture that combines Semantic and CBR
techniques to enhance the CCD process.
Our architecture is divided in five layers (see
Figure 1): (i) the User Layer, (ii) the Knowledge
Layer, (iii) the Experience Layer, (iv) the
Information Layer and (v) the Data Layer.
USING SEMANTIC TECHNOLOGIES AND CASE BASED REASONING TO SUPPORT COURSE CURRICULUM
DESIGN TASKS
385

From bottom up, the fist layer is the Data Layer.
This layer contains the data repositories that define
the different CCD conforming elements (e.g.
objectives, competences, courses, etc). The
components of the Data Layer are not necessarily
standardized; they are just bits of data that can be
used for a CCD.
DATA LAYER
(Disperse Data)
INFORMATION
LAYER
(Structured Data)
EXPERIENCE
LAYER
KNOWLEDGE
LAYER
CORE
COMPETENCES
ONTOLOGY
REASONER
ONTOLOGIES
eLearning
Standards &
Specifications
Query SystemMapperRepositories
CBR
(Reasoning)
CBR
(Data)
CORE
OBJETIVES
COURSES ………...
USER LAYER
Course CreatorTeacher ……
………...Objetives Courses Competences
Figure 1: Architecture.
The Information Layer (above) contains two
parts, (i) the ontologies and (ii) the CBRs.
The ontologies are constructed using a domain
model that can be fed from the Data Layer. This
means that these domain ontologies build their
individuals upon the data available in the first layer.
The CBRs are the second component of this
layer and they are a set of Case Systems (rules) that
use the data of the first layer as a feed.
It is interesting to note that the available data is
not necessarily used in both components and
moreover that some data collected is resultless
(revealing an opportunity for data model reduction).
At this level it can be argued that the data has
become usable information.
To convert the data into information a mapping
tool is needed. Such a mapping tool is implemented
based on the requirements of the domain model and
the CBR and it must be implemented inside both
components in a semi-automatic framework where
possible.
The information produced enters the Experience
Layer through a query system. Such a system
performs a series of queries over the ontologies and
the CBRs.
At the ontology level, the answers to the queries
are obtained using an ontology reasoner.
At the CBR level, the traditional CBR cycle acts
as the reasoner (based on rule logics and a statistical
analysis). Both the CBR and the ontology reasoner
are use to share information in a cyclic process.
When the information is processed it can be argued
that experience is obtained.
In the Knowledge Layer, the elements that
constitute the curricular plans are modelled, using e-
Learning standards.
Finally, in the User Layer several user types (e.g.
course creator, the teacher, etc.) are used to adapt the
system to particular cases.
4 CASE STUDY
Our case study is an application of the presented
architecture following the Spanish Ministry of
Education and Science (MEC) guidelines for
vocational education.
The domain was modelled based on the unit-
project composition of courses, following the
recommendations of our R&D project partner who is
a recognized expert in the field of CCD in Spain.
To explain the functionality of the prototype, it
must be mentioned that we implemented a use case
where the Course Designer user creates a course for
a non-existing competence .
The first task is to create a new competence
using the stored previous experiences (elements
contained in the Information Layer). For such
purposes, the user introduces the master guides of
the new competence that they are interested in.
Following these guides, the system launches a CBR
process on the data repositories containing the
competences. By doing so, they obtain a set of
similar competences stored in the repositories, which
are at this point in the Experience Layer. Those
competences will help them in new competence
generation. If necessary, it is possible to launch a
new CBR process changing the guidelines. When
finished, the new competence (that is now in the
Knowledge Layer) is saved (to the repositories of the
Data Layer) and becomes a part of the stored cases.
Once the new competence is created, the user
can decide if they want to assign a collection of
courses for the created competence, or finalize the
process and return to the competence creation task.
If the chosen option is to assign courses, an
ontology reasoner infers which are the most relevant
courses of the new competence based on a semantic
reasoning process performed over the stored courses
(these courses are in the Information Layer). With
this collection of suggested courses (and individual
units and projects, all of which are in the Experience
Layer), the user is able to design the new course. At
this point, the user can launch more CBR processes
CSEDU 2009 - International Conference on Computer Supported Education
386

to obtain new complete courses, or to obtain new
individual units and projects. The results obtained
are added to the course creation Experience Layer.
When the course contents are defined (in the
Knowledge Layer), it is necessary to establish a
common objective for the competence and the
course. The system shows the available objectives
(again from the Information Layer), and the user
chooses an available objective or creates a
completely new one. After this assignment, the new
course including its objective is stored in the
repositories of the Data Layer.
Finally, the user chooses between assigning other
course to the same competence or to end the process.
The User Layer filters the content displayed
depending on user type. In this use case, there is
only a user type, so there is not explicit
implementation of the layer.
4.1 Implementation Issues
The core language used to implement the prototype
was Java, using Swing and AWT libraries for
Graphical Interfaces. Competence, Course, Project
and Unit repositories were created and managed
with mySQL databases. The set of ontologies that
model the domain were written in OWL-DL, using
the Protégé ontology editor. For the query system,
we used Protégé OWL API (Knublauch, 2006), and
the chosen reasoner was Pellet (Sirin et al., 2007).
For the CBR implementation, we used jColibri2
(Díaz-Agudo et al., 2007), developed by the GAIA
group at Complutense University of Madrid.
5 CONCLUSIONS AND FUTURE
WORK
In this work, we presented an architecture to address
some common problems encountered in CCD.
Specifically we focused on the re-use of available
information. Our approach uses a mix of Semantic
and CBR techniques in order to enhance a real
world, factual industry problem. A case study
implementation of our architecture was presented for
using the design of a mid-level vocational education
course that complies with the Spanish normative as a
demonstration sample.
As future work we intend to extend our
implementation in two different directions, one
being related to the collaborative aspect of our work
(e.g. many users modifying the same resources at the
same time). The other direction we wish to explore
will focus on the possibility of enhancing the system
with experience in using SOEKS techniques (Sanin
et al., 2007) used in other domains with positive
results (Toro et al., 2007).
REFERENCES
Aamodt, A., Plaza, E., 1994. Case-Based Reasoning:
Foundational issues, methodological variations, and
system approaches. In AICom – Artificial Intelligence
Communications, 7 (1) pp. 39-59. IOS Press.
Diamond, R.M., 1998. Designing & Assessing Courses
and Curricula, Jossey-Bass. San Francisco.
Díaz-Agudo, B., González-Calero, P.A., Recio-García,
J.A., Sánchez, A., 2007. Building CBR systems with
jCOLIBRI. In Science of Computer Programming, 69
(1-3) pp. 68-75. Elsevier.
Fallon, C., Brown, S., 2003. e-Learning Standards. A
Guide to Purchasing, Developing, and Deploying
Standards-Conformant e-Learning, St. Lucie Press.
Boca Raton.
Gruber, T.R., 1995. Toward principles for the design of
ontologies used for knowledge sharing. In
International Journal of Human-Computer Studies, 43
(5-6) pp. 907-928.
Högskoleverket, 2008. E-Learning Quality: Aspects and
criteria for evaluation of e-leraning in higher
education, Report 2008:11 R, Swedish National
Agency for Higher Education.
Knublauch, H., 2006. The Protégé-OWL API. Webpage:
http://protege.stanford.edu/plugins/owl/api/index.html
(Last visited, 17 November 2008)
Leake, D., 1996. CBR in Context: The Present and Future.
In Case Based Reasoning: Experiences, Lessons and
Future Directions, pp. 3-30, AAI/MIT Press. Menlo
Park.
Noy, N.F., McGuiness, D.L., 2001. Ontology
Development 101: A Guide to Creating Your First
Ontology. In Stanford Medical Informatics Technical
Report SMI-2001, 880.
Rubio Oca, J., 2006. La Política Educativa y la Educación
Superior en México, 1995-2006: Un balance, Page
81,FCE. México.
Sanin, C., Szczerbicki, E., Toro, C., 2007. An OWL
Ontology of Set of Experience Knowledge Structure.
In Journal of Universal Computer Science, 13 (2) pp.
209-223, Graz University of Technology. Graz.
Sirin, E., Parsia, B., Grau, B.C., Kalyanpur, A., Katz, Y.,
2007. Pellet: A Practical OWL-DL Reasoner. In
Journal of Web Semantics, 5 (2) pp. 51-53, Elsevier.
Toro, C., Sanín, C., Vaquero, J., Posada, J., Szczerbicki,
E., 2007. Knowledge Based Industrial Maintenance
Using Portable Devices and Augmented Reality. In
KES 2007. Proceedings, Part I. LNCS 4692, pp. 295-
302. Springer Berlin. Heidelberg.
USING SEMANTIC TECHNOLOGIES AND CASE BASED REASONING TO SUPPORT COURSE CURRICULUM
DESIGN TASKS
387