2011 35th IEEE Annual Computer Software and Applications Conference Workshops
Finding Suitable Course Material through a Semantic Search Agent for Learning
Management Systems of Distance Education
Duygu ÇELİK, Computer Eng., İstanbul Aydin Univ., İstanbul/TURKEY
Atilla ELÇİ, Software Engineering Prog., Toros Univ., Mersin/TURKEY
3
Eray ELVERİCİ, Educational Tech., İstanbul Aydin Univ., İstanbul/TURKEY
1
2
1
duygucelik@aydin.edu.tr, 2atilla.elci@gmail.com, 3erayelverici@gmail.com
Recently, PowerPoint slides are recorded together with
sound and the finished product is converted to a flash file
using Adobe Presenter. Thus, asynchronous training media
content is prepared and presented. However, the components
of such systems can be strengthened in order to infuse more
intelligence into the system. For example, the system can
detect the levels and critical points of individual students and
provide feedback through an intelligent agent. We believe
that with such an approach it is possible to develop
contributions of the Semantic Web.
In this article, a system using Semantic Web
technologies is proposed in order to support learning
management system, to make such systems more efficient,
and to get meaningful results. In order to demonstrate the
proposed system work, necessary parts such as course
contents, course materials, and students’ profiles in IAUDEC
database are gathered to create ontology. For example,
Materials, Student Profile, Course and Exam ontology
knowledge bases; these ontologies are used in testing the
success of the system. Ontologies were created using Protégé
4.1 tool [3]. In addition, the proposed system and its core
module, the Semantic-based Course Material Browsing
Agent (CMBA), were develop on JAVA platform by using
JBuilder 2006 software [4].
Section 2 introduces theoretical backgrounds of this
paper. The system architecture, its management module, the
created ontology knowledge bases, and inferencing
mechanism are given in the third section. The third section
also gives example use cases and the results obtained.
Section 4 concludes this paper.
Abstract— Developing technology and increasing usage of
computers around the world make Internet a convenient tool
for Distance Education (DE) systems. Learning Management
Systems-(LMSs) of DE around the world serve content-based
courses to students/users via static pages on the Web.
Additionally, querying for specific course content of an entire
course in LMS is only possible by syntactic-based search
agents which limit the search. Therefore, it is thought that
querying of contents through semantic-based approaches
would make LMSs more helpful for a student/user through the
Semantic Web (SW) approach. Queries can be meaning based
and this can enable the system to make meaningful inference in
human-computer interaction. In this article, we propose a
system architecture that performs semantic–based searches
through an agent in order to plan student-based schedule of
course documents automatically according to student’s
performance in the course of a semester. OWL language of the
SW technology is used during the development of the ontology
of the LMS system that is used as a knowledgebase and
queried to obtain required data. In this article, we concentrate
on developing the system as a module for a LMS currently in
use.
Keywords; E-learning, Ontology, Semantic Web, Semantic
Resource Browsing Agent.
I.
INTRODUCTION
Distance education in the 1900s contributed to the
evolution of audio-visual devices. Today, distance education
practices have contributed to the development of media tools
[1]. The success of distance education depends on a
multitude of factors, a few of which are as follows: (1)
technology used, (2) quality components, (3) suitability of
LMS, (4) quality of multi-media based course content, (5)
corporate/institutional culture, and, (6) an effective feedback
mechanism. Distance education has long been used in
Turkey; one example of this is the practice of the Distance
Education Institute, Istanbul Aydin University (IAUDEC).
According to the research on DE systems, in order to
improve effectiveness of education, there should be more
than one DE component and components must support one
another [2]. These components may be synchronous training,
asynchronous training, exams, LMS-web pages, and distance
learning live support desk. High-tech education materials,
especially Adobe products are used for synchronous and
asynchronous training.
Today, course contents of distance education which are
presented as animations, graphics, videos and other multimedia materials are over PowerPoint presentations.
978-0-7695-4459-5/11 $26.00 © 2011 IEEE
DOI 10.1109/COMPSACW.2011.71
II.
SEMANTIC WEB
Semantic Web, which was developed as an extension of
today’s ordinary Web, was introduced by Tim Berners-Lee
and colleagues. Semantic Web can allow computers and
people work together by using well-defined meaning of
information [5]. For healthy functioning of Semantic Web,
computers should access to structurally organized data and
sets of inference rules which are used for automatic
reasoning. These sets are created as ontologies.
Ontology refers to the common meaning of knowledge of
the basic communication between people or systems which
is recognized by the relevant systems [6] detailing the
concepts in the field of information systems such as elearning [7 and 8]. In other words, ontology is the definition
of concepts and relations that an agent or a group of agents
may access [9]. Semantic Web requires a standard Web
ontology language to interpret the data meaningfully on the
386
Web environment. To meet this need, W3C had proposed
Web Ontology Language (OWL) in 2002, and, this language
became one of the W3C recommendations (i.e., standards) in
2004 [10]. OWL 2 was evolved recently [11].
The development of ontology knowledgebase containing
educational materials such as exams, slides, animations of
DE applications, and processing of such ontology by the
semantic-based Course Material Browsing Agent (CMBA)
are considered next. The CMBA, which is at the centre of the
system, is being introduced in detail in the next section.
III.
to interfere with the structure of the existing LMS system.
The benefit of the module is that it is designed as an internal
system which can be added in without breaking the interim
order and the mechanism of used LMS systems. The CMBA
System consists of the following components: User,
Distance Education Database (DE DB), Management
Module of Distance Education Database (MMDE-DB),
Ontology Knowledge bases (OKB), Inference Engine (IE),
and CMBA, the agent. The architectural structure of the
CMBA System is displayed in Figure 1; parts, objectives,
and working mechanism of the system will be discussed in
detail in this section.
ARCHITECTURAL STRUCTURE OF SEMANTIC-BASED
COURSE MATERIAL BROWSING AGENT (CMBA)
The architectural layout of the proposed system was
considered to be developed as an add-on module in order not
DE DB
5
4
DE DB: Distance Education Database
MMDE-DB: The Management Module of
Distance Education Database
SPT: Student Performance Table
CMBA: Semantic-based Course Material
Browsing Agent
IE: Infrence Engine
OKB:Ontology Knowledgebase
Student ID
2
Student/
Directors
Section
ORC
Performance
...
MMDE DB
ORC: Other Relational Courses
4
5
SPT
1
6
Taken Courses
Selected Course
ID
Student
Course
ID
6
3
CMBA
7
IE
7
OKB
Figure 1. Architectural structure of Semantic Source Browsing Agent
intended for students will not be mentioned due to the fact
that it is out of topic.) The aim here is to provide the system
with the ability of determining the standing of students in
terms of three levels as good/average/poor. The information
gathered will be shown as a table in the Management
Module of Distance Education Database (MMDE-DB) and
this Student Performance Table (SPT) will include the
student ID number, the selected course, units of the course,
achievement or performance score for the course, and the
other related courses, etc. (see step 4-5, Figure 1).
A. The Management Module of DE DB
CMBA System is thought to be used as an interface by
students and system directors alike. Namely, the CMBA
interacts with LMS through a user interface. Logging in with
a student ID card to the interface of the CMBA system, all of
the courses (in the LMS) that the student has been registered
for are seen.
In the first interface of the system, units of the student’s
lessons, exams and their results, the information about the
student’s term performance can be reached via the online
education database (see steps 1-3 above in Figure 1).
The information about the student and the results of level
evaluation are compared against a threshold value. The
performance score showing the student’s semester-based
performance is comprised of many parameters such as the
grades of his/her exams, the times of registration to a course,
attendance-absenteeism, the presented projects, assignments
and their results, the instructor’s marks, etc. At that point, the
system will take the necessary parameters required for
statistical calculation of an average achievement score. (In
this paper, statistical calculation of the achievement score
B. Ontology Knowledgebase (OKB)
Protégé 4.1 software was used to create the ontology
knowledgebase of the system [3]. The most focused ontology
is the Material.owl. Material ontology starts with 'owl:
Thing' class. Under this, the ‘Material’ class is defined as a
subclass. Four different subclasses are further defined under
the Material class. These are; book, presentation, test and
video (Table-1). As most of today's distance education
course materials are MS PowerPoint, Adobe PDF or MS
Word documents, and generally they are presentations, we
will deal with presentation section. However, in the created
387
Material ontology we can define all types of lesson materials
semantically.
and ObjectPropertyAssertion for some defined individuals is
given in Table 2.
TABLE I.
LABELS OF CLASS, DATATYPE, AND OBJECTTYPE
PROPERTY IN MATERIAL ONTOLOGY
TABLE II.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
46
THE CREATION AND DEFINITIONS OF THE MATERIAL OF
COURSE AND EXISTING UNITS
<?xml version="1.0"?>
Ontology xmlns=http://www.w3.org/2002/07/owl#
xml:base="http://aydin.edu.tr/Ontology/2011/Materİal.owl"
<Declaration>
<Class IRI="#Book"/>
</Declaration>
<Declaration>
<Class IRI="#Material"/>
</Declaration>
<Declaration>
<Class IRI="#Presentation"/>
</Declaration>
<Declaration>
<Class IRI="#Test"/>
</Declaration>
<Declaration>
<Class IRI="#Video"/>
</Declaration>
<Declaration>
<ObjectProperty IRI="#Covered_by"/>
</Declaration>
<Declaration>
<ObjectProperty IRI="#Covers"/>
</Declaration>
<Declaration>
<ObjectProperty IRI="#PreviousMaterial"/>
</Declaration>
<Declaration>
<ObjectProperty IRI="#NextMaterial"/>
</Declaration>
<Declaration>
<DataProperty IRI="#Interactive_Content_Property"/>
</Declaration>
<Declaration>
<DataProperty IRI="#Visual_Content_Property"/>
</Declaration>
<Declaration>
<DataProperty IRI="#Material_Address"/>
</Declaration>
<Declaration>
<DataProperty IRI="#Material_Language"/>
</Declaration>
<Declaration>
<DataProperty IRI="#Material_Author"/>
</Declaration>
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
<ClassAssertion>
<Class IRI="#Presentation"/>
<NamedIndividual IRI="#01_Math101_01_005"/>
</ClassAssertion>
<DataPropertyAssertion>
<DataProperty IRI="#Material_Address"/>
<NamedIndividual IRI="#01_Math101_01_005"/>
<Literal datatypeIRI="&xsd;URI">../lms/courses/math101</Literal>
</DataPropertyAssertion>
<DataPropertyAssertion>
<DataProperty IRI="#Material_Author"/>
<NamedIndividual IRI="#01_Math101_01_005"/>
<Literal datatypeIRI="&xsd;Name">Prof.Dr. Ali Gunes</Literal>
</DataPropertyAssertion>
<ObjectPropertyAssertion>
<ObjectProperty IRI="#Covered_by"/>
<NamedIndividual IRI="#01_Math101_01_005"/>
<Namedİndividual IRI=="#01_Cal102_01_001"/>
</ObjectPropertyAssertion>
<ObjectPropertyAssertion>
<ObjectProperty IRI="#NextMaterial"/>
<NamedIndividual IRI="#01_Math101_01_005"/>
<NamedIndividual IRI="#01_Math101_02_001"/>
</ObjectPropertyAssertion>
<DataPropertyAssertion>
<DataProperty IRI="#Question_Number"/>
<NamedIndividual IRI="#01_Math101_01_005"/>
<Literal datatypeIRI="&xsd;integer">12</Literal>
</DataPropertyAssertion>
x The unit of 01_Math101_01_005 is highlighted at line 3
in Table 2. It means the material numbered as 005 of the
freshman Mathematic course, Section/Unit 1 (i.e.
Grade_CourseName_Unit_MaterialID).
The
semantic
declaration (class assertion) of this material is defined as an
owl:individual in this ontology (see Table 2 line 1-4).
x We can see the semantic relations between the focused
course material 01_Math101_01_005 and the other course
materials. For instance, the material 01_Math101_01_005 is
Covered_by the material 01_Cal102_01_001 (see Table 2 at
lines 15-19).
x In DataProperty Assertions part, we can see that the
number of the questions of this material is 12 (see Table 2 at
lines 25-29).
x The author of it is Prof. Dr. Ali Gunes, and the level of it
is the 1st year (see Table 2 at lines 10-14).
x The address information is also present in order for this
material to be searched and found by a smart agent
developed semantically. The address of this material includes
an URL information which shows that it is in the Math101
file of LMS courses (see Table 2 at lines 5-9).
x The relation between the fifth material of the first grade
Math 101 course (01_Math101_01_005) and the second
unit material of the first grade Math 101 course
(01_Math101_02_001) is the next material. Namely, we
can define the materials required in sequence by using the
“#NextMaterial: ObjectProperty” (depicted in Table 2 at
lines 20-24).
We define common features of each material in this
ontology. We do this by using the feature definitions of
“ObjectProperty” and “DataProperty” that are present in
OWL language. The definitions of “ObjectProperty” that we
use in the ontology of materials are “#Covered_by”,
“#Covers”, “#PreviousMaterial and “#NextMaterial” (see
Table 1 at lines 20-30).
Additionally, definitions of “DataProperty” are:
“#Material_Address”,
“#Material_Language”,
“#Material_Author”,
“#Page_Number”
and
"#Course_Level" of the material, “#Question_Number” in
the materials, “#AnsweredQuestions_Number” in the
materials, and “#Multimedia_Content” (some of these are
depicted in Table 1 at lines 31-46). Furthermore, the
multimedia is comprised of five different subclasses:
interactive content, text content, visual content, image
content and audio content properties. A part of the material
ontology that showing examples of DataPropertyAssertion
388
C. Inference Engine
The SCA-Semantic Composition Agent-based System we
proposed in our earlier studies was about combining the
definitions of atomic work processes in order to constitute a
new and advanced process compatible with the query of a
user. In other words, it was aimed at a new complex work
process, which was not present but needed by a user, through
combining present process [13]. However, there is need for
suitable and correct inference rules in order to define a
semantic work flow.
The ‘Armstrong’s Axioms’ (AAs) seem to serve the
purpose. AAs form a set of axioms (inference rules) [12].
AAs are used in order to infer functional dependencies-FDs
that are encountered in a relational database. There are
seven different inference rules as seen in Table 3. Those are
Reflexivity, Augmentation, Transitivity, Pseudo Transitivity,
Additivity, Accumulation, and Projectivity.
In the light of such semantic information, a smart search
agent can perceive and interpret all features related to Math
101. In that way, it can decide on if a student needs this
material in terms of the information about the student’s
performance. Furthermore, the system can strengthen the
study program with present materials of related courses and
units.
Thanks to that, the smart search agents developed
semantically can detect the subsequent unit materials and
have the ability to make propositions when they are needed.
Similar to material ontology, course ontology and student
ontology have also been prepared in Protégé. However, in
this paper we concentrate on material ontology because we
focus on the role of a semantic search agent on the search for
necessary materials intended for students.
TABLE III.
ARMSTRONG'S AXIOMS AND REVISED ARMSTRONG’S AXIOMS.
Armstrong’s Axioms
1. Reflexivity: A set of attributes X determines a subset Y of itself: XoY if YX.
2. Augmentation: It allows enlarging the left-side of a FD or both side conventionally with one or more
attributes. Formally, if XoY then X||ZoY||Z for any Z.
3. Transitivity: If we have functions f : XoY and g : YoZ then we have a function g O f : XoZ, where
(g O f )(x) = g(f (x)). Formally, If {(X o Y) and (YoZ)},then XoZ.
4. Pseudo transitivity: is a generalization of Transitivity. It is requires the entire right hand side of a FD
appears as attribute(s) of the determinant of another FD. Formally, if {(XoY) and (Y||ZoW)} then
(X||ZoW).
5. Additivity: If there are two FDs with the same determinant on the left, it is possible to form a new FD
that preserves the determinant and has as its right-hand side the union of the right-hand sides of the two
FDs. Formally, if {(XoY) and (XoZ)} then (XoY||Z).
6. Accumulation: If there are two FDs with the complementary determinant, it is possible to form a new
FD that preserves the determinant X and forms its right-hand side as the union of the right-hand sides of the
two FDs except the complementary determinant Z. Formally, if {(XoY||Z) and (ZoC||W)} then
(XoY||C||W).
7. Projectivity: If X determines Y and Z, then X determines Y, therefore it is possible to break each
functional dependency XoY down to XoAi for i = 1..n where Y = {A1, . .An}. Formally, if (XoY||Z) then
XoY and XoZ.
In our earlier work [13], we used the AAs to detect the
relation between existing work processes in order to get new
work process. We proposed to merge the two inference rules
Reflexivity and Augmentation, so brought about Pseudo
Factorization axiom. Our aim with this single axiom was to
prevent derivation of excessive numbers and possibly
incompatible processes. In addition to this, we renamed
Projectivity axiom as Dissection. Thus obtained The Revised
Armstrong’s Axioms (RAAs) by transformations equivalent
to AA axioms, and during these transformations the
correctness and integrity of AA axioms were not disrupted.
In the first column of Table 4, we see AAs, and in the second
column we see RAAs (see Table 3).
Formal OWL meanings can provide various
characteristics of property. Those are transitive, symmetric,
functional, inverse of, inverse functional, etc. For instance,
if a property is transitive, for any X, Y, and Z materials the
property transitions of P(X, Y) and P(Y, Z) bring about the
meaning of P (X, Z). Kang et. al [14] proposed to monitor
and order the semantically defined work processes, namely,
to infer new property definitions among those work
Revised Armstrong’s Axioms
1. Pseudo factorization:
If {(X||YoZ) ,(T||ZoW)} and if (ZT||Z
and T≡X or T≡Y) then X||YoW.
2. Transitivity:
If {(X oY) and (Yo Z)} then (XoZ).
3. Pseudo transitivity:
If {(XoY) and (Y||ZoW)} then (X||ZoW).
4. Additivity:
If {(XoY) and (XoZ)} then (XoY||Z).
5. Accumulation:
If {(XoY||Z) and (ZoT||W)} then
(XoY||T||W).
6. Dissection:
If (XoY||Z) then XoY and XoZ.
processes, resorted to the definitions of owl:property present
for planning. However, in this study, we propose to use
RAAs to infer the property relations compatible among
materials, defining them as “owl: ObjectProperty. The
transitivity rule definition is shown in Table 4 (see the lines
1-5) and the usage of inference rules below for constitution
of relations among materials was planned as we prepared a
work plan. Therefore, in this work, RAAs have been applied
to the types of owl:ObjectProperty as mentioned before
(Covered_by, Covers, Previous and Next Materials) in order
to detect the order relations among materials (just as for
work processes) and constitute a material-based work plan
for student needs.
According to a scenario, a user enters into the LMS
system with his/her ID number and chooses the code of the
lesson for which a work plan is needed. LMS can reach all
information related to both the course and the student from
DE DB.
389
TABLE IV.
TABLE V.
OBJECT PROPERTY
VARIOUS MATERIAL RESOURCES
THE ONTOLOGICAL DEFINITION OF SOME RAAS
<!-- Object Properties of Tasks.Owl-->
1 <owl:ObjectProperty rdf:ID="hasTransitivity">
2
<rdf:type rdf:resource="&owl;TransitiveProperty"/>
3
<rdfs:domain rdf:resource="#Material"/>
4
<rdfs:range rdf:resource="# Material "/>
5 </owl:ObjectProperty>
ASSERTIONS ARE DEFINED FOR
1
2
3
4
5
<ObjectPropertyAssertion>
<ObjectProperty IRI="#Covered_by"/>
<NamedIndividual IRI="#01_Math101_01_005"/>
<NamedIndividual IRI="#01_Calc102_01_001"/>
</ObjectPropertyAssertion>
6 <owl:ObjectProperty rdf:ID="hasPseudoFactorization">
7
<rdf:type rdf:resource="&owl;TransitiveProperty"/>
8
<rdfs:domain rdf:resource="#Material"/>
9
<rdfs:range rdf:resource="# Material "/>
10 </owl:ObjectProperty>
6
7
8
9
10
<ObjectPropertyAssertion>
<ObjectProperty IRI="#NextMaterial"/>
<NamedIndividual IRI="#01_Math101_01_005"/>
<NamedIndividual IRI="#01_Math101_02_001"/>
</ObjectPropertyAssertion>
11 <owl:ObjectProperty rdf:ID="hasPseudoTransitivity">
12
<rdf:type rdf:resource="&owl;TransitiveProperty"/>
13
<rdfs:domain rdf:resource="#Material " />
14
<rdfs:range rdf:resource="#Material" />
15 </owl:ObjectProperty>
11
12
13
14
15
<ObjectPropertyAssertion>
<ObjectProperty IRI="#Covered_by "/>
<NamedIndividual IRI="#01_Math101_01_014"/>
<NamedIndividual IRI="#02_Math207_01_001"/>
</ObjectPropertyAssertion>
For instance, performance of the student is lower than the
minimum permissible value for Unit 1 of Math 101. LMS
searches for materials relevant to units which student has
shown inferior performance. Additionally, it will also
capture other related materials which are about the same unit.
Thus gathered is a repository of relevant materials.
In order to carry out inference operation, LMS consults
material.owl capturing ontology information on the relevant
material. As seen below in Figure 2, the property
‘NextMaterial’ of “owl:ObjectProperty” numbered 005
which belongs to Unit 1 of the first grade Math 101 is
connected to the material numbered 001 which belongs to
Unit 2 of the first grade Math 101 course. This relation is
associated semantically in Table 5, (see lines 6-10).
16
17
18
19
20
<ObjectPropertyAssertion>
<ObjectProperty IRI="#NextMaterial"/>
<NamedIndividual IRI="#01_Math101_02_001"/>
<NamedIndividual IRI="#01_Math207_01_004"/>
</ObjectPropertyAssertion>
21
22
23
24
25
<ObjectPropertyAssertion>
<ObjectProperty IRI="# Covered_by "/>
<NamedIndividual IRI="#01_Math101_02_001"/>
<NamedIndividual IRI="#02_Math207_01_001"/>
</ObjectPropertyAssertion>
The same semantic relation is defined between the material
numbered 001 which belongs to the Unit 2 of the first grade
Math 101 course and the material numbered 004 which
belongs to the Unit 1 of the first grade Math 207 course (see
Table 5 at lines 16-20). Therefore, the common units of two
courses and the content-based relation between them are
defined semantically.
Figure 2. ‘Next Material’ of “owl:ObjectProperty” feature defined for two materials (individual) in the Material.owl
The semantic material search agent will be able to
understand that after the material 01_Math101_01_005,
the material 01_Math101_02_001 needs to be processed,
and later bring about a work plan assuming that the
material 01_Math207_01_004 is necessary after the
material 01_Math101_02_001. For this, it will need a
transitive inference rule. It will use the transitive axiom: P
(X, Y) and P (Y, Z) imply P (X, Z) in order to detect the
material order.
ܔ܉ܑܚ܍ܜ܉ۻܜܠ܍ۼ
̴ܐܜ܉ۻ̴̴ ሳልልልልልልልሰ ̴ܐܜ܉ۻ̴̴
ܔ܉ܑܚ܍ܜ܉ۻܜܠ܍ۼ
̴ܐܜ܉ۻ̴̴ ሳልልልልልልልሰ ̴ܐܜ܉ۻૠ̴̴
then,
ܔ܉ܑܚ܍ܜ܉ۻܜܠ܍ۼ
̴ܐܜ܉ۻ̴̴ ሳልልልልልልልሰ ̴ܐܜ܉ۻૠ̴̴
390
Therefore, many axioms will give information about
the ordering among materials by being determined in the
suitable places in ontology, and make it possible to
comment on the work plan actively on the orderings.
Jena (JENA RDF API), one of the semantic web
technologies developed for semantic information inference
from ontology [15], and PELLET OWL DL Reasoner, one
of the semantic technologies for the reasoner operation
based on facts [16], are use commonly. Those SW
technologies are useful to extract data from ontology
knowledgebase and infer new facts from existing facts for
material order and quality. For instance, a student may
require a material that contains at least 20 solved
questions. Furthermore, a rank value about the materials is
computed based on various information such as number of
questions, number of answered example questions, the
person who prepared the material, quality of instruction,
picture, video, text or audio content by investigating the
“owl:DataProperty” properties. The quality of materials
searched and found, and the carriage of them to the upper
ranks in the work plans will increase the student
achievement and satisfaction. Thus, through defining unitbased relations among materials of courses in the ontology
files, and by doing necessary ordering and filtration for
constituting a qualitative work plan devoted to the needs
via inferences from these definitions, it is possible to
present the work plan needed by users.
The system can prepare and suggest a study plan
considering the student’s term performance, and it is
possible to comment on the student’s performance and
present a plan of materials if necessary. We need to
consider students term performance and define a good
study plan because DE-based LMS systems have a lot
registered students. This is crucial for such systems to
guide and assist their students who are normally out of
campus. This may be achieved through a smart module or
part of such system [13].
open for improvements and extension. In short, retrieving
right materials and producing a complex study plan were
made possible by using ontologies.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
IV.
CONCLUSION
In this article, we proposed a new system called
semantic-based Course Material Browsing Agent (CMBA)
which produces study plan by finding materials in an LMS
system semantically. We took the Online Education
Management System of Istanbul Aydın University as base
on which CMBA is grafted. Components of the offered
CMBA system are: User, Distance Education Database
(DE DB), Management Module of Distance Education
Database (MMDE-DB), Ontology Knowledgebase (OKB),
Inference Engine (IE) and CMBA the agent. CMBA has its
own inference rules; Revised Armstrong’s Axioms (RAA)
are used in planning and inference functions. In CMBA’s
inference module, RAAs are currently applied on only
objectType properties allowing the agent to perform
subject ordering. Besides, “owl:DataProperty” properties
are defined semantically in order to make quality choice
among materials. Several use cases are considered in
trying the developed system. The biggest advantage of this
system and the created ontology knowledgebase is that it is
[12]
[13]
[14]
[15]
[16]
391
The Institute for Higher Education Policy, Quality On The Line:
Benchmarks for Success in Internet-Based Distance Education.
Supported by National Education Association (NEA) and
Blackboard Inc. April 2000.
Hatanaka A, Okada A, Yuriyama M, Tarumi H, Kambayashi Y,
“Functions of Distance Learning System: VIEW Classroom,”
http://www.ineer.org/Events/ICEE2000/Proceedings/papers/MC21.pdfHatanaka, A., Okada, A., Yuriyama, M., Tarumi, H.,
Kambayashi, Y., Functions of Distance Learning System: VIEW
Classroom.
http://www.ineer.org/Events/ICEE2000/Proceedings/papers/MC21.pdf ,2000, last accessed November 2009.
Protégé, OWL-S Ontology Editor CS/AI Department, University
of Malta, 2004, http://owlseditor.semwebcentral.org/, last accessed
November 2010.
JBuilder
software
package,
Borland
JBuilder
2006,
http://www.borland.com/jb2006, last accessed November 2010.
T. Berners-Lee, J. Hendler, O. Lassila, The Semantic Web,
Scientific American, 284(5), 34-43, 2001.
Gruber, T. (n.d.). What is an ontology?. Retrieved June 15, 2007,
http://wwwksl.stanford.edu/kst/what-is-an-ontology.html,
last
accessed November 2010.
Ullrich, C., Description of an Instructional Ontology and its
Application in Web Services for Education. Proceedings of
Workshop on Applications of Semantic Web Technologies for Elearning SW-EL'04, 17-23, 2004.
A. Elçi, A Metadata Model For E-Learning Coordination Through
Semantic Web Languages, The Turkish Online Journal of
Educational Technology – TOJET, ISSN: 1303-6521 volume 4
Issue 3 Article 3, pp: 12-17, July 2005.
D. Çelik, A. Elçi, Provision of Semantic Web Services through an
Intelligent Semantic Web Service Finder, Multiagent and Grid
Systems - An International Journal, IOS Press, ISSN 1574-1702,
4(3) , 315-334, 2008.
OWL Web Ontology Language Overview, W3C Recommendation,
Online: http://www.w3.org/TR/owl-features/, last accessed
November 2010.
Grau, B.C. et. al. (2009) OWL 2-W3C Web Ontology Language.
W3C Recommendation by Oxford University April 2008-October
2009 http://www.w3.org/TR/owl2-primer/ Accessed: November
2010.
Armstrong, W. W., "Dependency Structures of Data Base
Relationships", Information Processing, 74, North Holland, 1974.
D. Çelik, A. Elçi, “Towards a semantic-based workflow model to
composition of OWL-S based atomic processes through process
similarity matching and inferencing techniques”, J. of Internet
Technology, ISSN: 1607-9264, Published by: Taiwan Academic
Network Executive Committee, 2010.
Kang, D., Lee, S., Kim, K., & Lee, J.Y., An OWL-based semantic
business process monitoring framework, Journal of Expert Systems
with Applications, 0957-4174, Vol. 36, Issue 4, 7576-7580, 2009.
Jena RDF API, Hewlett-Packard Development Company, 2001,
http://www.hpl.hp.com/semweb/jena.htm, last accessed November
2010.
E. Sirin, B. Parsia, PELLET - An OWL DL Reasoner, In
International Workshop on Description Logics (DL2004),
Whistler, Canada, 2004.