Content uploaded by Uday Bhaskar
Author content
All content in this area was uploaded by Uday Bhaskar
Content may be subject to copyright.
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 15
Adaptive Approaches to Context Aware Mobile Learning
Applications
Uday Bhaskar Nagella udaynagella@gmail.com
Research Scholar
Sri Venkateswara University
Tirupati, India
Dr. P. Govindarajulu pgovindarajulu@yahoo.com
Professor, Dept of Computer Science
Sri Venkateswara University
Tirupati, India
Abstract
Learning has gone through major changes from its inception in the human race.
Among all such major changes mobile learning is the latest to happen with the
advent of mobile learning technologies that have the potential to revolutionize
distance education by bringing the concept of anytime and anywhere to reality.
From the learner’s perceptive, mobile learning is “any sort of leaning that
happens when the learner is not at a fixed, pre-determined location or learning
that happens when the learner takes advantage of learning opportunities offered
by mobile technologies”. Research in context aware mobile learning has
concentrated on how to adapt applications to context. This paper reviews and
discusses few mobile learning systems the approach in implementing context
awareness and adaptation, and presents some others good work done in this
line.
Keywords:
adaptation, adaptive learning, context, learning activity, learner model, learning automata,
mobile learning.
1. INTRODUCTION
Mobile learning has become widespread, and students nowadays are able to learn anywhere and
at any time, enabled by mobile technologies and wireless internet connections. Mobile learning
can be distinguished “by rapid and continual changes of context, as the learner moves between
locations and encounters localized resources, services and co-learners” [1], and these different
situations are described by different learning contexts [2]. The diversity of mobile and wireless
technologies and the nature of dynamics in mobile environments complicate context awareness.
Context-aware mobile learning has become increasingly important because of the dynamic and
continually changing learning settings in the learner’s mobile learning environment giving rise to
many different learning contexts. The challenge is to exploit the changing environment with a new
class of learning applications that can adapt to dynamic learning situations accordingly. The task
of a context-aware mobile learning application is to sense the mobile environment and
react/adapt to the changing context during a student’s learning process [1]. Context is a key in the
design of more adaptive mobile learning systems [3] and context-awareness must be integrated
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 16
within the systems in order for them to be truly effective [4]. Mobile devices and sensing
technologies are combined to provide physical and environmental contexts in mobile applications.
Two characteristics of context have been described by Laerhoven [21] as activity and
environment. The task that the user is performing at the moment is described by the activity, and
focuses on the user of the device and his/her habits. The physical and social surroundings of the
user are described by the environment, such as the current location and movements in the
environment etc. The increasing use of mobile applications in varying contexts, locations and
surrounding environments means that if these applications were made context-aware, then
contextually relevant information from the devices can be transferred to the user [22].
In the mobile learning context, it is helpful to consider context awareness and adaptivity as two
sides of the same coin [5]. The purpose of the adaptivity and context awareness is to provide
better support for variety of learners, given that they may have very different skill and motivations
to learn in varying contexts. Adaptivity can be one form of adaptation; or as a quality of a system
to automatically and autonomically regulate and organize its functioning, the appearance of its
User Interface and the order of information it offers [6].
In this paper, we present some of the different approaches and methods to context aware and
adaptation/adaptivity that are already present in the literature. The remainder of this paper is
organized as follows. In Section 2, we describe the Characteristics of an ideal learning system
adaptive to the learner and to his/her context proposed by Telmo Zarraonandia, et al. In Section
3, we present three different (layered, modular, interaction) types of adaptive systems. In Section
4, we present different (Bayesian, Learning automata, Agent) type of methods for adaptivity
implementation. In Section 5, we present Bijective adaptation between context and adaptivity in a
mobile and collaborative learning. Finally, conclusions and future work is given in Section 6.
2. DESIRABLE CHARACTERISTICS OF AN IDEAL LEARNING SYSTEM
ADAPTIVE TO THE LEARNER AND TO HIS/HER CONTEXT
Telmo Zarraonandia, Camino Fernandez, Paloma Diaz & Jorge Torres[7] have selected a group
of desirable characteristics for an ideal learning system that is able to adapt the course to learner
specific characteristics, knowledge, objectives and learning goals, also sensitive to the context in
which the learning session is taking place and capable to adjust the appropriate parameters
accordingly.
Different kinds of information has been separated and represented by means of different models
like Domain Model, User Model and Adaptive Model, of which Adaptive Model is of our interest
here.
“Adaptive Model: This model relates the two other models (Domain & User) defining which
materials will be presented and how the presentation will take place in order to achieve the
learning goal, optimizing the learning process. The characteristics that were considered for this
model are:
• To use some kind of pedagogical approach for the adaptation.
• Pedagogical rules updateable: As the pedagogical approach may vary depending on
the learning theory that is applied or on the knowledge domain, they wanted their
system to be flexible enough to give the instructors the possibility of programming
different pedagogical approaches for the same course.
• Levels of sequencing of learning material: to adapt the sequence of concepts
presented to the user in order to achieve his/her learning goal (knowledge routes) and
to adapt the sequence of learning materials presented to the user in order to achieve
the knowledge of a particular concept (content sequencing)..
• Learner progress consideration: Capability to generate the knowledge routes of
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 17
materials dynamically by taking into consideration the learner progress during the
course.
• Re-routing: Capability to re-plan the knowledge route if special difficulties are detected
with a particular concept and even to change the current pedagogical method if the
system detects it is not performing well for that particular learner.
• Use of standards to define the rules that govern the adaptation.”
The above said characteristics are of most important when the learning system transforms into a
mobile learning system taking into account all the dimensions of mobility.
3.
ADAPTIVE SYSTEMS
3.1 Context Aware and Adaptive Learning
In architecture [8] of a context-aware and adaptive learning, Jane Yau and Mike Joy proposed
two Adaptation Layers: The information has been considered for the survey purpose and
presented here as taken from the cited reference.
“Learning Preferences Adaptation Layer: This layer consists of the Learning Preferences
Adaptation Module, which contains sub-modules - Learning Styles Adaptation, Learning Priorities
Adaptation and Knowledge Level Adaptation.
The learning preferences of a learner are retrieved from the Learner Profile database and
incorporated into the relevant sub-module for the appropriate learning objects to be chosen at a
later stage.
Matching the correct level of information according to the learner’s most appropriate learning style
can also create a more enjoyable and effective learning experience for the learner [9].
Contextual Features Adaptation Layer: This layer consists of the Contextual Features Adaptation
Module, which contains sub-modules – Location-specific Adaptation and Time-specific
Adaptation. Each of the contextual features are retrieved from the Learner Schedule database
and incorporated into the relevant sub-module for the appropriate learning objects to be chosen
at a later stage.
By placing a User Verification option, the problem of learner not conforming to his/her schedule is
rectified; This prompts the user at the beginning of the learning session to indicate whether the
location and the available time that the tool has retrieved is accurate. Another method which can
also be used to detect discrepancies between the learner’s stated location and his/her current
location is to have an option for Software Verification.
The importance of obtaining the actual location of the user derives from the fact that the
contextual features surrounding the location are different in various different places, and can
affect the learner’s ability to study such as their concentration level, which can be affected by the
level of noise in the location or environment.
A number of methods for obtaining the noise level to determine the possible level of concentration
that the learner has at a location have been considered. Firstly, a microphone sensor can be
used to detect the noise level which can approximately indicate the level of concentration that the
learner has in such an environment with that level of noise. Secondly, results obtained by Cui and
Bull [10] can be used to map the concentration level of a learner to a certain type of location.
In the study conducted, student’s location of study and their corresponding chosen level of
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 18
concentration were recorded and discovered that the chosen concentration levels in various types
of location by different students were found consistent even though noise levels may have been
different. The results indicate that students shared similar levels of concentration in the same
location despite the varying levels of noise. This also suggests that there are other elements in
the environment which could affect a student’s concentration level such a movement.”
3.2 Context - based Adaptation in M-Learning
In architecture [11] for a Context-based Adaptation in M-Learning proposed by Estefania Martin,
Nuria Andueza, Rosa M. Carro manages data about users and activities so that the most suitable
activities to be accomplished at each time are suggested to each user are taken care of by
Activity adaptation module.
“Activity adaptation module: This module is responsible for deciding about the availability of
activities and of generating the list of available activities to be sent to the alert module which is
responsible for alerting the learner about the same depending on the various contextual element
values and by processing a set of rules that indicate in which cases the situation of the user is
appropriate for alerting him/her about the availability of activity.
This module is structured into three sub-modules: Structure-based adaptation, Context-based
General Adaptation and Individual Adaptation.
Firstly, the Structure-based Adaptation sub-module processes the structural rules, which
establish, for each type of user accessing the system, the relationship between activities, as well
as the order in which they must be performed, if any. Its main aim is to generate a list of activities
to be suggested to the user. The first step is to select the activities according tot the most
appropriate rules for a certain user. For the same activity, the list of sub-activities can be different
depending on certain conditions related with the user’s personal features, preferences, as well as
his/her current situation, including the context(spare time, location, available devices), pending
activities and actions during his/her interaction. Therefore, the rule activation conditions are
checked and the corresponding rules are triggered.
Secondly, the Context-based General Adaptation sub-module consists of a filter that processes a
set of general rules to choose the type of activities more suitable of being accomplished by the
user. It adds/removes activities depending on their type (review, individual exercises,
collaborative activities or messages, among others). This filter affects to all the activities to be
performed.
Finally, the Individual Adaptation sub-module checks the conditions of atomic activities, if any.
These conditions can be related to any user feature or action stored in the user model.”
Rule-based adaptation techniques are used in the above discussed system.
3.3. Interaction through Context Adaptation
In the investigative study [12] by Yuan-Kai Wang shows importance of context-awareness and
adaptation in mobile learning, proposes Context Aware Mobile Learning (CAML) that senses
mobile environment and reacts or adapts to changing context during learning process has four
interaction modes.
The challenge is to exploit the changing environment with a new class of learning applications
that can adapt to dynamic learning situations accordingly. Interaction through situated or reactive
adaptation can improve learning process. There are four key modes of interactions in Context
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 19
Aware Mobile Learning (CAML):
“(a) Spatio-temporal dependent interface: It is situated user interface adapted according to time
and location contexts. For a mobile learner in classroom at course period, lecture slides and
student notes are most important interfaces. However, homework and group discussions become
primary when the learner changes place to home after course period. Located learning objects
that are nearby or meaningful are emphasized or otherwise made easier to choose.
(b) Contextual event notification: Learning process is mostly planed as a calendar with a lot of
scheduled activities, such as lecturing, test, examination, homework, and so forth. Timely
execution of some course activities, such as the reminding of homework, can be implemented as
context-triggered event. Notification is dynamically scaled and adapted by inferring interruptibility
based on the user’s recent activity and context of the mobile environment. The interruptibility of
event notification could be spatio-temporal context dependent as a simple example. Facility and
activity contexts are also helpful for contextual event notification.
(c) Context-aware communication: Communication can be divided into asynchronous and
synchronous messaging between teachers and students, or among students. Asynchronous
messaging, such as email, discussion board are desired when the recipient is unavailable or if
either is not currently near a computer. Synchronous messaging, such as online chats are more
appropriate after course for group discussion. Context of online status can be used to gauge
whether the learner is in a course context or a social context where an interruption is less
appropriate. Spatio-temporal, facility and activity contexts are important for the appropriate
utilization of communication methods.
(d) Navigation and retrieval of learning materials: Learner can reactively or actively browse and
search learning materials. In reactive learning, accurate learning materials are delivered to the
learner if the activity context of personal learning progress is obtained. In active learning, effective
browsing and searching of tremendous learning materials are important and can be achieved by
context restriction. For example, proximate selection is one way of context restriction by spatial
context.”
4. IMPLEMENTING ADAPTIVITY IN MOBILE LEARNING
4.1. Using Learning automata as probabilistic adaptation engines
Economides A.A. proposed adaptation engine in an Adaptive Mobile System [13] that used
Learning automata to implement the probabilistic adaptation decisions.
“Adaptation Engine: The inputs to the Adaptation engine are – learner’s state, the educational
activity’s state, the infrastructure’s state and the environment’s state (Table1). The output
consists of the adapted educational activity’s state, and the adapted infrastructure’s state (Table
1).
U(t)=[L(t),A(t),I(t),E(t)] is the input to the adaptation engine at time t.
O(t+1)=[A(t+1),I(t+1)] is the output from the adaptation engine at time t+1.
Input U(t) Output O(t+1)
L(t): Learner’s state
A(t):Educational
Activity’s state
I(t): Infrastructure’s state
E(t): Environment’s state
A(t+1):Adapted
educational Activity
I(t+1):Adapted
Infrastructure.
TABLE 1
:
Input and Output of the Adaptation Engine
.
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 20
Learning Automata Adaptation: Here probabilistic algorithms to adaptively select the most
appropriate state of the educational activity or/and the infrastructure are proposed. They employ
learning automata that reinforce a good decision and penalize a bad one [14].
At time t, the adaptation engine selects the state for the educational activity to be A(t)=Am with
probability PAm(t), and the state for the infrastructure to be I(t)=tn with probability PIn(t). Define
PA(t)=[PA1(t),…,PAM(t)], and PI(t)=[PI1(t),…,PIN(t)].
Considering Learning Automata Adaptation decisions, following are given:
Assume that at time t, the A(t)= Am is selected probabilistically according to PA(t).
If this results in “good” outcome (e.g., the learner is satisfied) then increase the probability of
selecting again the Am and decrease the probabilities of selecting all other As.
Otherwise, do the opposite.
Assume that at time t, the I(t)= In is selected probabilistically according to PI(t).
If this results in “good” outcome (e.g., the learner is satisfied) then increase the probability of
selecting again the In and decrease the probabilities of selecting all other Is.
Otherwise, do the opposite.
For example, Let assume that there are two networks in the vicinity of the mobile learner. The
problem is to select the network that will provide her the best communication performance and
reliability in order to achieve her educational activity.
Therefore, let I1 be the Infrastructure including the first network, and I2 be the Infrastructure
including the second network.
Let also, PI1 be the probability of selecting the I1, and PI2 be the probability of selecting the I2.
Let at time t, In(n=1 or 2) is selected with probability PIn(t).
If the communication performance and reliability delivered to the learner is “good”, then increase
PIn(t+1), the probability of selecting again infrastructure In: PIn(t+1)=PIn(t)+a*(1-PIn(t)), 0<a<1,
Otherwise, decrease PIn(t+1): PIn(t+1)=PIn(t)-b*PIn(t), 0<b<1,
Of course, PI1(t+1) +PI2(t+1) =1.
In the above example, the Linear Reward-Penalty learning automation has been used. However,
other learning automata algorithms [14] may also be used depending on the situation.”
4.2. Using Bayesian network to determine mobile learner’s style and adapt to it
Yu Dan and Chen XinMeng [15] used the Bayesian networks to determine mobile learner’s styles
exploring the potential of individualization of learning process for the learners to implement
adaptive mobile learning system architecture that provides a mechanism for adapting content
presentation to the mobile learner model and device model, improving mobile learning process.
“Each individual has his/her unique way of learning. Learning style greatly affects the learning
process, and therefore the outcome[16]. Mobile learners, who are typically distance learners,
usually work individually without external support and have various learning backgrounds and
levels. This work here uses Bayesian network to determine mobile learner’s styles, which is
based on Felder-Silverman learning style theory.
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 21
Felder-Silverman learning style theory categorizes an individual’s preferred learning style by five
dimensions: active/reflective, sensory/intuitive, visual/auditory, sequential/global and
inductive/deductive [17]. As inductive/deductive dimension has been deleted from the previous
theory because of pedagogical reasons, here they modeled four dimensions of Felder-Silverman
framework in their application domain. They built a Bayesian network representing the learning
style with a knowledge engineering approach[18].
For each dimension they analyzed respective determining elements in mobile environment, and
listed these elements and their values that can take in the following:
Active/reflective (Processing):
Wiki: participation, no participation.
Short message reply: many, few.
Forum: reply, browse, no use.
Sensory/intuitive(Perception):
Reading: facts, theory.
Example: before exposition, after exposition.
Visual/auditory(Input):
Learning material: audio, video.
Chat: audio, video.
Sequential/global(Understanding):
Information Processing: step by step, jump.
Answer: result after steps, only result.
According to the above analyses, they implemented by Bayesian network(Fig.1) encoding
relations among three types of variables: learning styles, four dimensions of the learning styles,
and different elements that determine learning styles.
FIGURE.1
Bayesian network modeling mobile learner's styles
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 22
Once establishing the probability values associated with each node of the graph by expert
knowledge and collected data, they made probabilistic inference of a learner’s style. For example,
suppose for a learner, we obtain the probability values from observation for the reading and
example, and then with the conditional probability table of node Perception they compute the
probability P (perception) to determine the student is a sensory learner or an intuitive learner.”
4.3. Supporting Adaptivity in Agent-Based Learning Systems
Shanghua Sun, Mike Joy & Nathan Griffiths developed an agent-based learning system [19] that
incorporates learning objects to facilitate personalization, and is based on a learning style theory
as the pedagogic foundation for adaptivity, and evaluation indicates that the approach is able to
provide personalized( or adapted) learning materials and improve the adaptivity in learning
systems.
“Here the system locates the student’s learning style preference into the learning style space, and
also stores each student’s current learning style, and the style attributes of each learning object,
as coordinates in the four-dimensional space. The system will then search the repository of
learning objects, to fetch appropriate learning object with similar dimensional descriptions. These
are supported by agent technology to realize the algorithm and implement the process. The
objects are then presented to the student, and the subsequent interactions between the student
and these learning objects may be used to modify the learning style attributes recorded for a
student.
Agent technology has been used to facilitate autonomy and adaptivity, decoupled from the
pedagogic foundations of the system [20]. Their system consists of five agents as shown in the
figure 2, namely Student Agent, Record Agent, Modelling Agent, Learning Object Agent and the
Evaluation Agent. Each agent is designed to satisfy a certain functional requirement to actualize
the service purpose of the education system, namely to provide dynamic and adaptive learning
materials to individual users.
FIGURE 2
.
System Architecture
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 23
The Evaluation Agent ensures that learning objects are presented in individual and adaptive
learning paths to each individual student. Here the use of learning objects and learning style in a
agent-based learning system to enhance adaptivity has been introduced. At the conceptual level,
personalization and adaptivity are achieved by the use of learning style schemes to tailor the
presentation of learning objects to individual students. Conversely, at the practical level, this
adaptivity is achieved by providing a set of agents that use a combination of prebuilt and acquired
knowledge to determine the leaning styles and learning objects that are appropriate for individual
students. Evaluation of the system effectiveness and efficiency is to be assessed.”
5. Bijective Adaptation between context and activity
Jihen Malek, Mona Laroussi and Alain derycke [3] have presented an innovative approach for
modeling a Bijective adaptation between context and learning activities within mobile and
collaborative learning environments; in which is an adaptor that defines two classes of
functionalities: the adaptation of learning activity to context and the adaptation of context to
learning activity.
“Adaptor for Bijective Adaptation: This adaptor models all possible interactions between context
and learning activity because context and learning activities influence each others in learning
processes. This adaptor provides developers with all the possible adaptation actions that should
be taken into account through a bijective adaptation process.
i. Activity Adaptation to Context
An activity is adapted or conducted with some variations depending on some values of contextual
elements attributes. So, only relevant context changes are modeled and taken into account for
the triggering of the activity adaptation process.
When event (or relevant context changes) occurs, conditions (or adaptation rules) are checked
and then Activity adaptation process actions are triggered.
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 24
Three levels are differentiated within an activity adaptation process:
Presentation level: for example, if the physical environment is characterized by a high noise level
and the learning activity includes a sound, then adaptation action compels the inhibition of this
sound if it is not necessary for performing the activity.
Navigational level: to model the navigational schema of learning activity, they used UML activity
diagrams as shown in figure 3 below. Adaptation actions at this level will consist of the selection
of the appropriate learning sub-activity according to the current context.
FIGURE 3.
Navigational schema of learning activity
For example, the number of connected learners of the teamwork is used as selection criteria
which outlay the nature of the adequate sub-activity.
Intentional level: The intentional level doesn’t have the same role as the other levels of
adaptation. This level guarantees that learning activity’s objectives are preserved and not
modified by the adaptation whatever the latter is; presentation or navigational level.
ii. Context Adaptation to Activity
a. Activity updates context
The values of some contextual element’s attributes are updated when an activity is completed,
but some conditions must be checked before the update (when activity’s goals are achieved).
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 25
For example, passing an exam updates the user model in a way that extends the learner’s
knowledge in that relative area.
b. Activity adapts context
Some contextual elements can be adapted to the activity needs. So context adaptation process
consists of controlling the context parameters in order to adapt them to the activity needs. This
adaptation aims to create an adequate learning environment which helps learners concentrate
better on their learning activity.
For example, the volume of the tape recorder stereo will be adjusted automatically according to
the type of activity performed by the learner and the degree of needed concentration.”
6. Conclusion and Future Work
This paper presents few mobile learning systems approach in implementing context awareness
and adaptation. Adaptation is achieved in terms of the learner context, the learner’s knowledge
levels, the content that is to be presented, the learner’s style which may vary from learner to
learner and Adaptation to the device’s context. Finally we identify that there is a need for explicitly
modeling the entities in a mobile learning environment including learner and his behavior in
dynamically varying contextual elements and representing inherently existing associations
between all the above said approaches to be modeled for an effective mobile learning system
which enhances the mobile learner experience, knowledge and usage of the device. We conclude
that further research is needed in modeling the user interactions at application level, activity level
and at user interface level to understand user intentions and acceptability levels for a given
context(s) to make mobile learning more intelligent for adaptation and hence become learner
friendly resulting in maximization of the mobile learning goals.
7. REFERENCES
[1] Chan, T., Sharples, M., Vavoula, G. and Lonsdale, P. (2004) Educational Metadata for mobile
learning, International Workshop on Wireless and Mobile Technologies in Education.
[2] Derntl, M. and Hummel, K. (2005) Modelling context-aware e-learning scenarios. Pervasive
computing and communications workshop.
[3] Malek, J., Laroussi, M. and Derycke, A. (2006) A Multi-Layer Ubiquitous Middleware for
Bijective Adaptation between Context and Activity in a Mobile and Collaborative learning.
ICSNC 2006.
[4] Lavoie, M. (2006) Mlearning: Identifying Design Recommendations for a context-aware
mobile learning system. IADIS International Conference Mobile Learning.
[5] Antti Syvanen, Russell Beale, Mike Sharples, Mikko Ahonen and Peter Lonsdale. (2005)
Supporting Pervasive Learning environments: Adaptibility and Context Awareness in Mobile
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 26
Learning, In the Proceedings of the 2005 IEEE International Workshop on Wireless and
Mobile Technologies in Education (WMTE’05).
[6] Anu Jappinen, Mikko Ahonen, Teija Vainio, Erika Tanhua -Piiroinen.(2004). Adaptive mobile
learning systems – the essential issues from the design perspective. In the proceedings of
MLearn 2004.
[7] Telmo Zarraonandia, Camino Fernandez, Paloma Diaz & Jorge Torres (2005). On the way of
an ideal learning system adaptive to the learner and her context, In the Proceedings of Fifth
IEEE International Conference on Advanced Learning technologies (ICALT’05).
[8] Jane Yau and Mike Joy (2007). Architecture of a Context-aware and Adaptive Learning
Schedule for Learning Java. ICALT ,2007.
[9] R. Beale and P. Lonsdale, Mobile Context Aware Systems: the intelligence to support tasks
and effectively utilize resources, Mobile HCI, 2004.
[10]Y. Cui and S. Bull, Context and learner modeling for the mobile foreign language learner,
Science Direct, System 33, pp.353-367, 2005.
[11]Estefania Martin, Nuria Andueza, Rosa M. Carro (2006), Architecture of a System for
Context-based Adaptation in M-Learning, In the Proceedings of the Sixth International
Conference on Adavanced Learning Technologies (ICALT’06).
[12]Yuan-Kai Wang (2004), Context Awareness and Adaptation in Mobile Learning, In the
Proceedings of the second IEEE International Workshop on Wireless and Mobile
Technologies in Education (WMTE’04).
[13] Economides, A.A.(2006), Adaptive Mobile Learning, Proceedings WMUTE – 4th International
Workshop on Wireless, Mobile and Ubiquitous Technologies in Education, 2006.
[14] Economides, A.A., Multiple response learning automata, IEEE Transactions on Systems,
Man and Cybernetics, Vol. 26, No 1, pp.153-156, February 1996.
[15]Yu Dan, Chen XinMeng, Using Bayesian Networks to Implement Adaptivity in Mobile
Learning, Proceedings of the second International Conference on Semantics, Knowledge,
and Grid(SKG’06).
[16]Kinshuk and T. Lin, Application of Learning Styles Adaptivity in Mobile Learning
Environments, in Third Pan-Commonwealth Forum on Open Learning, Dunedin, New
Zealand, 2004.
[17] R. Felder and L. Silverman, Learning and Teaching Styles, Journal of Engineering Education,
vol.78, no.7, 1988,pp.674-681.
[18] P. Garcia, A. Amandi, S. Schiaffino and M. Campo, Using Bayesian Networks to Detect
Students’ Learning Styles in a Web-based Education System, in Proc of ASAI, Rosario, 2005,
pp. 115-126.
[19] Shanghua Sun, Mike Joy & Nathan Griffiths, To Support Adaptivity in Agent-Based Learning
System – The Use of Learning Objects and Learning Style, Proceedings of the fifth IEEE
International Conference on Advanced Learning Technologies (ICALT’05).
[20] S. Sun, M. Joy, and N. Griffiths, The use of learning objects and learning styles in a multi-
agent education system, Proc. of ED-MEDIA 2005.
Uday Bhaskar Nagella & Govindarajulu
International Journal of Computer Science and Security, Volume (2) : Issue (2) 27
[21] K.Laerhoven, “Online adaptive context-awareness starting from low-level sensors”, PhD
Thesis, 1999.
[22] J. Kolari, T. Laakko, T. Hiltunen, V. Ikonen, M. Kulju, R. Suihkonen, S. Toivonen and T.
Virtanen, “Context-aware services for mobile users –
technology and experiences”, VTT publications 539, 2004.
