Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

An Intelligent eTutor-Student Adaptive Interaction Framework

Shehab A. Gamalel-Din
King Abdulaziz University, Computing Information Systems, Jeddah, KSA
drshehabg@yahoo.com

Abstract However, researchers have guidelines for effective and efficient


Many educators believe that the most effective means for teaching education. Snow and Farr [1] suggested that sound learning
is through one-on-one interactions with students. This research’s theories are incomplete or unrealistic if they do not include a
hypothesis is that better learning results would be achieved by whole person view, integrating both cognitive and affective
adapting the e-tutor interaction with its individual student user. aspects, implying that no educational program can be successful
eTutor-Student interaction in this research is based on adapting without due attention to the personal learning needs of individual
the content and presentation of the learning material to the student students. Russell [2] suggested that educators should identify and
based on his/her learning model—The student model. In other acknowledge learning differences and make maximum use of the
words, eTutor should adapt and personalize the teaching strategy available technology to serve them accordingly. Brusilovsky [3]
for each student; something that is not easy to achieve without the suggested using adaptive hypermedia to support individual
aid of an intelligent system with a comprehensive knowledgebase. learning. The idea of adaptive e-learning is to adapt both the
This article presents one essential component of our research on content and presentation of the course based on the profile of the
adaptive e-learning—namely, a framework of a Smart Cognitive learner. Paolucci [4] addressed the importance of
Augmented Learning Object Repository (SCALOR) engine that individualization in that any strategy should be adaptive and
augments the concepts of learning styles onto Hypermedia personalized. To ensure personalization, adaptive systems should
Learning Objects, which together with a Smart domain be capable of diagnosing and identifying each student’s
knowledge ontology compose the Smart e-Learning misconceptions. Based on this conception, Gamalel-Din [5] have
Knowledgebase (SELK). SELK is at the core of the developed a framework for an adaptive web-based intelligent
personalization of the eTutor-Student interaction for a more tutoring system employing a student model that is developed by
efficient and effective learning process. Evaluation results for Gamalel-Din [6].
this framework proved the hypothesis. Therefore, one of the most formidable tasks for educators is
shaping their presentations of core knowledge to meet the needs
Categories and Subject Descriptors of each individual learners having varied and diverse cognitive
H.1.2 [Models and Principles]: User/Machine Systems - human and psychological traits [7]. This research tried to achieve such a
factors, human information processing, software psychology. goal through introducing Smart e-Learning Knowledgebase
(SELK), which is the core heart of the adaptation process of the
General Terms learning material personalization to each individual learner.
Human Factors. SELK has three pillar that are materialized onto three
components: first, a detailed model of the individual student--the
Keywords Smart Student Model (SSM); second, learning materials must be
eTutor-Student interaction, Student model, Cognitive model,
composed of small granule multimedia objects referred to as
Learning style, Domain ontology, Adaptive e-Learning.
atomic learning objects (ALOs), third, domain knowledge
1. Introduction ontology that interlinks both SSM and ALO for an efficient and
Educated and skilled human resources and workers are becoming effective adaptation process.
keys of success and power of both nations and organizations. The student model is used for tailoring the teaching strategy and
Therefore, education and training are expected to have a more dynamically adapting it according to the abilities of each
important role than ever before. Specifically, effective and individual student. Users’ Models are often based on various
efficient Web-based intelligent adaptive e-learning systems are different dimensions. Brusilovsky [3] defines these Dimensions
argued to be solutions for several challenges facing education. as: users’ goals, knowledge, background, hyperspace experience
and preferences [8]. The focus of this paper is on two of those
dimensions, namely, background knowledge and learning style.
A learning style is defined, among many definitions, as “the
unique collection of individual skills and preferences that affect
how a student perceives, gathers, and process learning materials”
[9]. In section 2, a more detailed description on SSM is discussed
together with how it personalizes the eTutor-Student interaction.
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies On the other hand, Learning objects are drawn from repositories
are not made or distributed for profit or commercial advantage and that (LOR) and are specified using standard metadata formats, such as
copies bear this notice and the full citation on the first page. To copy
SCORM and IEEE LOM. This research suggests extending the
otherwise, or republish, to post on servers or to redistribute to lists,
requires prior specific permission and/or a fee. LO metadata to accommodate new attributes that are utilized by
Interaccion’12, Oct 3–5, 2012, Elche, Alicante, Spain. the Smart e-Learning engine to augment cognitive characteristics,
Copyright 2012 ACM 978-1-4503-1314-8/12/10...$10.00 especially Learning Styles. These repositories are called Smart
Cognitive Augmented Learning Object Repository (SCALOR).
The appropriate LOs are drawn from SCALOR according to
specific rules, which are driven by this research, in such a way describes two possible architectures for both building SELK and
that most suits each individual student’s model: learning style and delivering personalized learning material. Section 4 presents
background knowledge. The LO selection is based on proper experimental evaluation of the model. Section 5 reviews
identification of the appropriate values of metadata attributes literature and presents some related work. Finally, Section 6
specifying the required material. concludes.
The Knowledge domain is represented by an ontology. Domain 2. The Smart e-Learning Knowledgebase
knowledge ontology is a semantic network of nodes representing
concepts that are interrelated by relationships with different types. (SELK)
This research has extended the ontology representation into what This section presents the three components of SELK in more
is called Smart Ontology Model (SOM) that is a multilayered details. The abstraction of each component representation is first
network whose concepts nodes are complex structures with levels presented and then followed by a discussion on how it is utilized
corresponding to those of Bloom’s instructional development in adapting the course presentation to individual students.
taxonomy [10]. SOM is interlinked to SCALOR to interrelate
ALOs to concepts.
2.1 Student’s Smart Conceptual Background
Model – Smart Ontology Model (SOM)
SELK is considered by the 12-years-long Smart e-Learning Domain knowledge are represented in structured Ontology that
research project “Smart School of the Future” [11,12] a central consist of concepts and relationships among these concepts as
component to all processes throughout the lifecycle of e-learning, shown in Figure 3a. The background knowledge of each student is
as depicted by Figure 1. Smart course authoring and delivery assessed and is projected on the Domain Ontology model to
environments as well as assessment utilize SELK for smart generate an individualized Student’s conceptual background
personalization. In addition, many education support tools were knowledge ontology model that is a subset of the general domain
designed around the concepts of learning model to reveal better knowledge ontology.
results [13,14,15,16,17] as shown in Figure 2.
Revised Bloom’s
SO1 SO1 Concept Levels
F 6 Remembering
S
u
p

[1]
[4 t T
p

Concept FConcept
] oo
o

Authoring Understanding
r

1 Concept 1
6
ls

Applying
Has Part Has Part
Has Part Has Part Analyzing
Cognition- Prerequisite M M Applying_
Require Evaluating
Augmented SELK Concept Follows Concept Concept Follows Concept SO1.2
SO1.1 SO1.1
2 SO1.2 2 Creating
3 3
Revised Bloom’s
Levels
A

[2] Has Part Has Part Has Part Has Part


ss

[3 sm

Remembering
es

] e

Delivery
D D Understanding
n
t

Concept Concept Concept Concept Applying


4 5 4 5 Analyzing
Figure 1: Adapted SELK is Central to the whole e-Learning Evaluating
SO1.1.1 SO1.1.2 SO1.1.1 SO1.1.2
Lifecycle Creating

a) Traditional Domain Ontology b) The Smart Ontology Model

Figure 3. Abstract Example of Domain Concepts’ Ontology

In domain ontology, concepts are interconnected by a set of


semantic relationships, such as:
 HP (Has Part): X (Y1, Y2,….Yn) means that, the concept X is
composed of or can be covered by the concepts Y1, Y2, ...,
Yn. It is interpreted as “Y1, Y2, ..., and Yn are detailed topics
of X that can be covered when teaching all or part of these
subtopics depending on the required depth.”
 F (Follows): F (X, Y) means that it is preferable to teach Y
and X in this order. This relation supports the concept
ordering when designing a course.
 R (Related): R (X, Y) means that, the concepts X and Y are
Figure 2. Smart e-Learning Tools and Environments under related, i.e., conceptually one of those subtopics is sufficient
“Smart School of the Future” to cover the parent concept.
Section 2 discuss the different constituent components of SELK
discussing how learning styles and background knowledge are  P (Prerequisite): P (X, Y) means that, the concept Y requires
urgently required to personalize eTutor-Student interaction and the concept X, i.e., before you can teach X, you need to
demonstrating how it improves the e-learning process Section 3 assure that Y is already captured by the learner. This
relation determines if the student is missing a certain the input dimension (visual/verbal) and the understanding
background X in order to be able to follow on a specific dimension (sequential/global). Each student has a preference on
concept Y, if so, a compensating topic or preview is added to each of the four dimensions, e.g., sensing/ visual
the one-on-one course delivery. /active/sequential). Figure 5 depicts the four dimensions of the
FSLSM.
The Smart Ontology Model (SOM), as shown in Figure 3.b, has
two improvements on the traditional model described above. In
SOM, the concept’s node is a complex structure. Each node is
given a complexity value (F=Fundamental | M=Medium |
D=Advanced) that is used to guide the design of a training course
according to the course’s complexity. For instance, if the course
is a medium-level course, all advanced concepts (Marked with D)
can be ignored; only F & M concepts are included.
The second improvement in SOM is the accommodation of the a) Student1’s BK Model
Revised Bloom’s Taxonomy (RBT) [18]. Each concept node is
made of six levels corresponding to Bloom’s levels. This will
make SOM as a multilayered diagram; one layer for each of the
Bloom’s levels. This improvement is intended to guide the course
design phase in which the course objectives specify the target
Bloom’s level for each concept covered in the course.
Accordingly, this concept’s SOM’s layer is employed and the
relationship links are followed. Most importantly of those links is
the prerequisite link which might reference a specific layer of
another concept, as shown in Figure 3b. where Concept5 is b) Student2’s BK Model
having complexity “M” and whose “Applying” level requires
Student1 Student2
Concept6 at Level “Understanding”. 1. Problem Representation
The P relationship is utilized in adapting the course content in 2. AI Search Strategies
2.1. Introduction to AI
case of a missing prerequisite. Assuming that concept Y requires
Search strategies
concept X as a prerequisite at a certain RBT level L, then the 2.2. List of AI Search
Smart Adaptive course preparation/delivery system should use the strategies
SOM to assure that the course plan accommodates the missing 2.3. Illustration of AI
knowledge of X before Y is presented. Figure 4 shows how the Search strategies
course design is adapted according to the SOM of two different 3. Uninformed Search 1. Uninformed Search
students. It is assumed in this example that Student1 hasn’t 3.1. Definition of 1.1. Definition of
studied neither “Problem Representation” nor “Search Strategies”, Uninformed Search Uninformed Search
Strategies Strategies
which are prerequisites for all types of search, while student2 has
3.2. List of Uninformed 1.2. List of Uninformed
acceptable applying level in both concepts. Therefore, the course Search Strategies Search Strategies
content is adapted to accommodate what Student 1is missing. 3.3. Description of 1.3. Description of
Uninformed Search Uninformed Search
2.2 Student’s Learning Styles Strategies Strategies
Each individual has his/her unique way of learning. Learning 4. Depth First Search 2. Depth First Search
style greatly affects the learning process, and therefore its 4.1. 2.1.
outcomes [9]. The underlying idea of a learning style approach is c) Adapted Course’s Topics for Each Individual Student (Course
that a person learns more effectively when information is Table of Contents)
presented in a manner that matches his/her preferred methods of Figure 4. An Example of Course Adaptation based on
acquiring and processing information. each Students’ Background Knowledge
Learning styles are defined differently by different researches. For
example, Alonso [19] defined learning styles according to Definition Dimension Definitions
cognitive psychology as “personal manners to perceive and Do it Active Reflective Think about it
process information, and how they interact and respond to
educational stimuli”, while Keefe [20] defined learning styles as Learn facts Sensing Intuitive Learning
“characteristic cognitive, affective and psychological behaviors concepts
that serve as relatively stable indicators of how learners perceive, Require Visual Verbal Require reading
interact with and respond to the learning environment”. Pictures or lecture

Recently, there arose many learning style theories, such as: Kolb Step by step Sequential Global Big picture
learning style theory [21], Gardner’s Multiple Intelligences
theory, and Felder-Silverman learning style theory. FSLSM [22] From Inductive Deductive From
is the learning styles model that is adopted by SELK as it is more particulars to generalities to
comprehensive and leaning towards adaptation more. Figure 5 generalities particulars
sheds lights on the different dimensions of FSLSM. FSLSM
incorporates four dimensions, the perception dimension Figure 5. Felder-Silverman’s Learning Style Model [22]
(sensing/intuitive), the processing dimension (active/reflective),
SELK utilizes the learning style model in choosing the most When the students registered to the system, they were asked to fill
appropriate pedagogy that suite the best for each individual in a student model assessment questionnaire the purpose of which
student. The learning style directs both the selection and is to identify the values of each parameter of SELK for each
sequencing of the learning objects (LOs) when delivering a course individual student. Those parameter values compose the specific
material to a student, as depicted in Figure 6, while Table 1 student model for each individual, which will aid in adapting the
summarizes how the pedagogy is managed. material presentation to the student accordingly.

Visual vs. Verbal


Selection

Sequential vs. Global


Selection &
Sensing vs. Intuitive Sequencing

Active vs. Reflective Sequencing

Figure 6. The Effect of Learning Styles on Organizing


Learning Materials
a) Course Framework for Sequential and Global Students
Figure 7 depicts two examples to demonstrate how SELK adapts
the presentation according to Students’ LS. Moreover, Figure 8
demonstrates how LS directs student assessment by suggesting
where exercises should be located. Notice that the position of the
exercises section changes according to the learning style.
This research suggested modifying and adapting the LO metadata
standards by adding extra attributes as well as employing some
other attributes after stretching their space of acceptable values in
such a way that is necessary for supporting the theories
accommodated by SELK.

3. The Delivery Engine Architecture


Figure 9 depicts the architecture of the SELK building and
maintaining environment, while the architecture of the Smart e-
Learning course delivery environment is shown in Figure 10. The
Delivery engine is responsible for providing adaptive courses
personalized for each individual student according to his/her
learning model. The student learning model is identified and b) Course Framework for Intuitive and Sensing Students
saved at the student registration time through a LS questionnaire
and Knowledge assessment tests. This model is then used all over Figure 7. An Example demonstrating how SELK Adapts
the student interaction with the Smart Tutor. Course Delivery According to Student’s LS

4. Evaluation
In order to validate SELK, a prototype was developed using
Microsoft .NET Visual Basic (VB) and Active Server Pages
(ASP), which allows the dynamic generation of pages with Visual
Studio 2005 and executed on a Windows XP/Vista platform. A
course module was also implemented in the domain of Artificial
Intelligence specifically Depth first and Breadth first search as
uninformed search strategies.
One of the methods followed in evaluating the model was to
conduct two experiments in each of which three groups of
uniformly distributed students, of a total of thirty students, were
formed according to Graf’s study [32]. The students were
distributed to the three groups randomly such that their GPAs are Figure 8. Positions of exercises based on student’s learning
uniformly distributed across the groups to become styles – Sample Personalized TOCs
probabilistically equivalent.
Table 1. Recommended Implications of FSLSM on Pedagogy

Style Selection Criteria Sequencing Strategies

 Present the general overview LO at the beginning of each chapter or


section.
Retrieve the LOs that have Depth type = {General
Global  The sequence of the general overview LOs depend on their
Overview, In-depth}.
presentation order in the original materials.
 Finally, present the in-depth LOs of each section.
Sequential Retrieve the LOs that have Depth type = { In-depth}. Present the in-depth LOs in their sequence in the original materials.
Retrieve the LOs that have Concrete content  Content Present the LO with the Example Instructional Format before the LO that
Sensing
type = “concrete”. have Explanation Instructional Format.
Retrieve the LOs that have Abstract content  Content Present the LO with the Explanation Instructional Format before the LO
Intuitive
type = “Abstract”. that have Example Instructional Format.
Present the Exercises immediately after the related LO or the related
Active −−−
section.
Present the Exercises at the end of each chapter (sensing) or at the end of
Reflective −−−
the course/e-book (intuitive).
Retrieve LO that have the Format = {Figure, Graph,
Visual −−−
Image, Video Clip}.
Retrieve LO that have the Format = {Table, Text, Audio
Verbal −−−
Clip}.

Table 2. LO’s Extended Metadata Attributes as suggested by SELK

Attribute Description Values

Concept Concept covered by this LO


Technical {Pictures, Graphs, Images, Diagrams, Flowcharts, Schematics, Concepts maps, Animation,
Technical data type for LO
format Video, Audio, Schematics, Text, Highlighted text, Hypertext}.
{Introduction, Overview, Definition, Fact, Remark, Example, Explanation, Description,
Instructional Illustration, Comparison, Summary, Conclusion, Theory, Rule, Formula, Procedure, Algorithm,
Specific kind of LO
role Exercises, Case study, Real world problem, Question , AnswerToQuestion}.

Cognitive That is covered by this LO according


{Remembering, understanding, applying}
level to RBT
Content type Type of concept {Concrete, Abstract}.
Teaching
Teaching strategy used within the LO {Expository Explanation, Inquisitor Explanation, Assessment}.
Strategy

Students belonging to the first group (referred to as matched


Table 3: The Results of the First Experiment
group) were presented with a course that matched their learning
styles. The second group (referred to as mismatched group) got a
Matched Mismatched Control
course that mismatched their learning styles. The third group
Group Group Group
(referred to as control group) was provided with a course where
all available learning objects were presented in a default sequence M SD M SD M SD
independent of the students’ learning styles. After studying the
material through the system, students were given a post-test for Time spent to study the
18 5.37 30 6.24 27 11
assessing their learning outcomes and performance. course (in minutes)

To assess the efficiency of the learning process, the first Score of the post-test
8.8 1.62 7.30 3.68 8.75 3
experiment put no time limit on the students to finish studying the (10 marks)
material; once the student finishes studying he/she were presented
with an assessment quiz. While the second experiment was
designed to assess the effectiveness of the learning process by Table 4: The Results of the Second Experiment
limiting the study time to a maximum of 30 min. for the student to
finish studying the material after which they were presented with Match Mismatch Control
the evaluation quiz to assess their depth of understanding. In both Group Group Group
experiments the system measured the elapsed time per each Mean Of post test Score 7.05 5.14 5.36
student. Table 3 describes the results of the first experiment
while Table 4 shows the results of the second experiment.
The results clearly demonstrate that the model presented by SELK
enhanced both the efficiency and effectiveness of the learning
outcomes of the students.
Book Ontology
Book Ontology
Builder

LO Extractor Incremental General Domain Ontology


Ontology Builder

Instructor
Instructor
LO Metadata Builder LO Relationship Builder
Central Knowledge-
Base Generator

LO Metadata LOR Model LO Repository


attributes

Learning Style Knowledge


Identifier StudentModel
Student Model
Identifier

Background
Knowledge

Learning Styles
Student Model
Student Figure 10.Identifier
The Architecture of eTutor-Student Adaptive Interaction Delivery Engine
Student

Figure 9. SELK Building Architecture

Smart
Ontology
Model

Figure 10. The Architecture of eTutor-Student Adaptive Interaction Delivery Engine


5. Related Work knowledge, learning style, skills and traits model, and emotional
Adaptive educational hypermedia is an inspiring area of research mode. This article discusses the use of the first two components
that combines Intelligent Touring Systems (ITS) with educational only for personalizing the learning material presentation. SOM is
hypermedia [3] employing the concept of student model. Two a especially designed comprehensive ontology structure to
types of adaptation arose, namely, adaptive presentation and accommodate instructional theories, such as RBT, to aid both the
adaptive navigation, which proved positive effects on student's instructor and the student during the two phases of course
learning and comprehension [23]. Adaptive presentation systems, development and delivery, respectively.
such as InterBook [24], and AHA [25], adapt the content of the In order to verify and evaluate SELK and its constituents,
learning materials, while adaptive navigation systems, such as prototypes were implemented, experimental lectures were
ELM-ART [25], and WEAR [26], adapt the hyperspace links to designed, and experiments were conducted on computer science
help students to find their optimal paths. students. Results revealed that such a model enhances both the
Student model is a core part of all adaptive educational efficiency and effectiveness of the learning process and achieves
hypermedia systems, and so is the work in this reseaech. The improved outcomes.
different aspects of the student model that are mostly used in The bigger research project “Smart School of the Future” is
adapting the content and navigation of the hypermedia are: goal, continuing its efforts in extending SELK to investigate other
experience, preferences, background knowledge, and learning dimensions of the student model, such as thinking style, cognitive
styles. This article focuses on the latter two while a more model, IQ, objectives…etc. In addition, the project is continuing
comprehensive model is condidered by the bigger 12-years-long the investigation of the impact of such components on the
Smart e-Learning project [11,12]. The field of learning styles has learning process on a larger scale. A plan is set to apply the
been incorporated in few prototypes of educational systems to model on a full scale course rather than a single lecture. This full
satisfy the students’ differences in perceiving and processing course will be used as a basis for evaluating the whole model and
information; examples are CS383 [27], LSAS, INSPIRE, Tango its impact on both the efficiency and effectiveness of the learning
[26], [28]. process. On the other hand, such a project should also evaluate the
On the other hand, educational metadata records are developed to effort and cost required to develop courses according to the
extend the scope of standard metadata by adding additional fields, concept set by SELK. In addition, it must identify the kind of
which describes information that has particular educational support an instructor would require of tools as aids for designing
relevance [29,30,31]. This research, likewise, suggested courses according to SELK’s model.
modifying and adapting the LO metadata standards by adding
extra attributes as well as employing some other attributes after
7. References
stretching their space of acceptable values in such a way that is [1] Snow, R, and Farr, M.1977. Cognitive-conative-affective
necessary for supporting the theories we accommodated. processes in aptitude, learning, and instruction: An
introduction, Conative and affective process analysis, 3.
6. Conclusion [2] Russell, T. 1977. Technology wars: Winners and losers,
Educational and cognitive theories showed that in one-on-one Educom Review, 32(2).
learning sessions, better outcomes would be achieved if learner’s
[3] Brusilovsky, P. 2002. Developing adaptive educational
models were properly used in designing and delivering learning
hypermedia systems: from design models to authoring tools,
material. Instructor’s submitted material should be presented in
in T. Murray, S. Blessing and S. Ainsworth (Eds.)
the form of a set of LOs augmented with suitable metadata
Authoring Tools for Advanced Technology Learning
attributes upon which personalization will depend. The 12-years-
Environments: Toward Cost-Effective Adaptive,
long Smart e-Learning research project “Smart School of the
Interactive, and Intelligent Educational Software, Ablex,
Future” has demonstrated how student model can be utilized in
Norwood
order to personalize both the presentation of learning material to
each individual learner. For instance, the learner background [4] Paolucci, R. 1998. Hypermedia and learning: The
knowledge interferes with the course design for this specific relationship of cognitive style and knowledge structure, In
student to cover those missing concepts but in such a way that the Proceedings of ED-MEDIA/ED-TELECOM, Freiburg,
matches the theme of this specific course: goals and complexity. Germany, June 20–25, 1998.
Furthermore, the learning style of each individual learner would [5] Gamalel-Din, S. 2002. The Smart Tutor: Student-Centered
drive both the selection and sequencing of cognitively augmented Case-Based Adaptive Intelligent e-Tutoring, in the
LOs. On the other hand LO selection should be directed by a Proceedings of the First International Conference on
comprehensive domain ontology model. Informatics and Systems, INFOS2002.
This article presents a framework of the Smart e-Learning [6] Gamalel-Din, S. 2002, Have We forgotten Emotions?
Knowledgebase (SELK) that is a core constituent of the 12-years- Temporal Proactive Student Modeling, AUEJ, Vol 5, No. 4,
long Smart e-Learning research project “Smart School of the Sept. 2002.
Future”. SELK integrates three components: Smart Student [7] Whitehurst, R., Powell, C. and Izatt, J. 1998.Utilizing the
Model (SSM), Smart Cognitive Augmented Learning Object Student Model in Distance Learning, In the Proceedings of
Repository (SCALOR), and Smart Ontology Model (SOM). the ITiCSE’ 98, ACM.
SCALOR extends the structure of metadata attributes of learning
objects (LO) to augment cognitive characteristics so as to to allow [8] Brown, E., Brailsford, T., Fisher, T., Moore, A. and
for personalizing both the content and presentation of the learning Ashman, H. 2006. Reappraising Cognitive Styles in
material according to each individual smart learner’s model Adaptive Web Applications, In the Proceedings of the
SSM. SSM extends the concept of student model to International World Wide Web Conference Committee
accommodate four components of a learner’s profile: background (IW3C2), ACM, Edinburgh, Scotland, May 23–26.
[9] Hong, H., and Kinshuk, 2004. Adaptation to student In the Proceeding of Second IEEE International Conference
learning styles in web based educational systems, In L. on Advanced Learning Technologies (ICALT'01), IEEE,
Cantoni & C. McLoughlin (Eds.), Proceedings of ED- Madison, WI, USA, pp 185-188.
MEDIA 2004 - World Conference on Educational [24] Brusilovsky, P., Eklund, J., and Schwarz, E. 1998. Web-
Multimedia, Hypermedia & Telecommunications (Lugano, based education for all: a tool for development adaptive
Switzerland), June 21-26. courseware, In the Proceedings of the seventh international
[10] Anderson, L. and Krathwohl, D. 2001. A Taxonomy for conference on World Wide Web 7, Elsevier Science,
learning teaching and assessing: A revision of Bloom's Brisbane, Australia, pp 291 – 300.
taxonomy of educational objectives. Addison Wesley [25] Henze, N. 2000. Adaptive Hyperbooks: Adaptation for
Longman, Inc, New York. Project-Based Learning Resources, Ph.D. Thesis, University
[11] Gamalel-Din, S. 2009. Smart E-Learning School of the of Hanover, Germany.
Future: Project Report”, In the Proceedings of the 6th E- [26] Stash, N., Cristea, A., and Bra, P. 2004. Authoring of
learning Applications Conference, Cairo, 10 -12 Jan. learning styles in adaptive hypermedia: problems and
[12] Gamalel-Din, S. 2009, Smart E-Learning: A Greater solutions, In the Proceedings of the 13th international World
Perspective, From the 4th to the 5th Generation e-Learning, Wide Web conference on Alternate track papers & posters,
In the Proceedings of the First International Conference on International World Wide Web Conference, New York, NY,
Operations Research and Applied Informatics (CORAI 09), USA, pp 114 – 123.
Cairo, Egypt, 15 – 19 February. [27] Carver, C., Howard, R. and Lane, W. 1999. Enhancing
[13] Gamalel-Din, S., Al-Saad, F. 2008, Smart E-NoteBook: an student learning through hypermedia courseware and
Adaptive Hypermedia Learning Material Management incorporation of student learning styles, IEEE Transactions
Environment, The 7th European Conference on e-Learning, on Education, Volume 42, Issue 1, pp 33-38, Feb.
Cyprus , 6-7 November 2008. [28] Graf, S., Kinshuk, and Kappel, G. 2007. Adaptivity in
[14] Gamalel-Din S., and Al-Saad, F. 2008, Learning styles, Learning Management Systems Focusing on Learning
Learning Objects, and Adaptive e-Learning, Al-Azhar Styles, Ph.D. Thesis, Vienna University of Technology,
Engineering 10th International Conference, Cairo, 24-26 Faculty of Informatics, Vienna, December.
December, 2008. [29] H.S. Al-Khalifa and H. C. Davis, 2006. The Evolution of
[15] Noaman, A., Gamalel-Din, S., Essa, F., and Al-Sherbini, M. Metadata from Standards to Semantics in E-Learning
2008, Building an E-Learning Content Management Applications, In the Proceedings of the Seventeenth
Framework, Journal of Administrative research, July 2008. Conference on Hypertext and hypermedia.
[16] Al-Ayed, A., and Gamalel-Din, S.2009. The Smart Office [30] Bourda, Y. and Doan, B. 2003. The Semantic Web for
Hours Assistant: an Intelligent Student-Centered FAQ Learning Resources, In the proceeding of the 3rd IEEE
System, In the Proceedings of ECEL 2009, The 8th International Conference on Advanced Learning
European Conference on e-Learning, Bari, Italy, 29-30 Technologies.
October 2009 [31] McClelland, M. 2003. Metadata Standards for Educational
[17] Gamalel-Din, S., Al-Otaibi, R. 2010. Intelligent Querying Resources, Volume 36, Issue 11, Nov.
For Adaptive Course Preparation and Delivery in E- [32] Graf, S. and Kinshuk, K. (2007) Providing Adaptive
Learning, In the Proceedings of the Ninth IASTED Courses in Learning Management Systems with Respect to
International Conference on Web-based Education, Sharm Learning Styles. In Proceedings of World Conference on E-
El-Sheikh, 15-17 March, 2010. Learning in Corporate, Government, Healthcare, and Higher
[18] Anderson. L. and Krathwohl, D. 2001, A Taxonomy for Education, AACE, 2007, pp. 2576-2583.
learning teaching and assessing: A revision of Bloom's
taxonomy of educational objectives, Addison Wesley
Longman, Inc, New York.
[19] Alonso, C. 1993. Educational technology and learning
styles: Rethinking the Roles of Technology in Education, In
the Proceedings of the Tenth International Conference on
Technology and Education, pp 1277-1279.
[20] Keefe, J. 1979. Learning style: an overview”, In J.W.
Keefe, editor, Student learning styles: diagnosing and
prescribing programs. NASSP, 1979.
[21] Mustaro, P., Silveira, I. 2006. Learning Objects: Adaptive
Retrieval through Learning Styles, Interdisciplinary Journal
of Knowledge and Learning Objects, Volume 2.
[22] Felder, R. and. Brent, R. 2005. Understanding Student
Differences, Journal of Engineering Education, Vol. 94 No.
1.
[23] Moundridou, M., and Virvou, M. 2001. Authoring and
Delivering Adaptive Web-Based Textbooks Using WEAR,

You might also like