43

International Journal of Learning, Teaching and Educational Research

Vol. 17, No. 7, pp. 43-58, July 2018
https://doi.org/10.26803/ijlter.17.7.3

Affect-Adaptive Activities in a Personalised

Ubiquitous Learning System

O. S. Adewale, O. C. Agbonifo, E. O. Ibam, A. I. Makinde,

O. K. Boyinbode, B. A. Ojokoh and O. Olabode
Department of Computer Science, School of Computing, The
Federal University of Technology, Akure, Nigeria

M. S. Omirin and S. O. Olatunji

Ekiti State University, Ado – Ekiti, Nigeria

Abstract. Many literatures have shown the importance of emotion in

learning because of their effect on learner‟s performances thereby giving
reasons why learners‟ affective states are crucial to learning. In this light,
this research work aims to identify affect-adaptive activities that would
improve learning in a personalised ubiquitous learning system by
detecting learning style and affective states through some patterns of
behaviours. In the study, learners‟ preference are determined, learners
affective states such as confidence, effort, independence and confusion
are investigated. The ant colony clustering algorithm is used to
determine learners‟ activities. The four affective states are determined
from the learners‟ forum threads, content activities, assessment, timing,
and discussion through learners‟ engagement. The result of this research
on two separate courses, Course A (Affect) and Course B (Non-Affect)
shows that the four affective states influence learning. The average mean
and standard deviation (SD) value for Course A (Mean=276.23,
SD=272.27) over Course B (Mean=200.58, SD=210.51) showed an
improvement in learning performance and the t-Test carried out
between the courses suggested that students‟ performance is dependent
on their affective states.

Keywords: affective states; adaptive; ubiquitous; learning; learning

styles; activities.

1. Introduction
The numerous developmental phases of information technology have
transformed education system in recent years. This development gave access
to global communications and resources available to today's students at all
levels of schooling (Yaun et al., 2013; Jones & Jo, 2004).

© 2018 The authors and IJLTER.ORG. All rights reserved.

 44

Learning styles make it possible to assign the preferred ways of learning.

Whenever a learner with a unique learning style is not supported by any
teaching environment, it makes learning difficult for such a learner (Kirschner
et al., 2006; Felder & Silverman, 1988). For instance, learner learns more when
the instruction presented correspond to their learning style (Entwistle &
Ramsden, 2015). The best method for getting learning style which is based on
instructional material in large classes is through adaptive e-learning
environment as this does not have the common limitations face to face
scenario has which focus on individual students due to the lack of required
resources and time (Wanner & Palmer, 2015).

The application of ubiquitous computing in education known as ubiquitous

learning (u-learning) marks a great step forward in the delivery of education
instruction materials. The u-learning has revolutionized education and
overcome some of the constraints of traditional learning. However, learning
using the same learning content will lead to a cognitive mismatch. The
integration of adaptive learning and ubiquitous computing may offer great
innovation in the delivery of learning, granting access to personalization and
customization to student needs (Sung et al., 2016; Kim, 2006).

Furthermore, in a traditional learning environment, a teacher finds it difficult

to teach all learners‟ individually due to a large number of the learners.
However, the teacher can monitor and draw conclusions from students‟
learning behaviour which can be referred to as learners‟ affective states. A
learning system that considers learner‟s affective state will improve and
therefore enhances a learner‟s learning experience (Baker et al., 2010).

Several parameters such as anger, fear, joy, surprise, disgust, motivation,

interest, proclivity, confidence, confusion, and effort have been used to
describe learners‟ affective states (Khan et al., 2010; Qu et al., 2005;
Lopatovska & Arapakis, 2011). Affect has been shown to influence the
learning process (Kort et al. 2001; D‟Mello et al. 2014; Baker et al. 2010).
Affective states can be classified into two, the positive affective states and the
negative affective states. The positive affective states such as confidence,
satisfaction and so on contribute towards learnings while the negative
affective states inhibit learning. Therefore, the paper investigates the effect of
adaptivity, personalisation, ubiquity and affective states on students‟
learning.

2. Literature Review
With the emergence of ubiquitous learning (u-learning), access to education
by students have become flexible, and seamless. The definition of u-learning
can be broadly classified into three categories: the adopted technologies, the
education objects and the learning process (Peña-Ayala & Cárdenas, 2016). In
a definition influenced by the classification of learning environments, u-
learning was described as a combination of the e-learning and m-learning
paradigms (Behera, 2013).

© 2018 The authors and IJLTER.ORG. All rights reserved.

 45

Learning Styles
A learner whose learning style is not supported by a particular learning
environment is bound to face difficulties during learning (Entwistle &
Ramsden, 2015; Felder and Silverman, 1988). Felder and Silverman Learning
Style Model (FSLSM) is one of the learning style models used in learning style
related research for information technology-based learning systems. The
FSLSM has been demonstrated to be most appropriate for hypermedia
courseware (Tortorella & Graf, 2017; Liyanage et al., 2016; Fasihuddin et al.,
2017). FSLSM distinguishes among four different dimensions. These
dimensions are active, reflective, sensing, intuitive, visual, verbal, sequential
and global dimension. A combination of these dimensions makes up the
learners‟ individuals learning preferences.

Advances in information technology have transformed greatly the practice of

learning in recent years. However, learning platforms such as Moodle,
Blackboard among others do not support a full-fledged adaptivity that is
based on learning styles and affective states (Jones & Jo, 2004).

Adaptive Learning
Adaptive learning can be referred to a concept that learning methodologies to
learners' individuals learning styles. Some of the elements of adaptive
learning has been demonstrated to include: monitoring activity, interpreting
the results, understanding requirements and preferences, and facilitated
learning process with gained information (Zimmerman, 2013; Paramythis &
Loidl-Reisinger, 2004).

This type of learning system offers a great advantage in providing learners

with specific and personalised knowledge during learning. The use of
adaptive strategies can be dated back as early as the start of the twentieth
century as educators began to look for other means to improve the learning
outcomes of student understanding and learning (Moore & Kearsley, 2011;
Zimmerman, 2013). The development of such systems will definitely provide
or support effective learning experiences for a different type of learners across
a broader view of knowledge domains. Therefore, adaptive learning systems
are to feature learner preferences, interests, learning contents, and browsing
behaviours to provide adaptive and personalised learning services
(Premlatha, 2016; Wang et al., 2017; Beldagli & Adiguzel, 2010).

Personalised Learning
Personalised learning allows learners‟ specific requirements to be identified
and addressed during a learning process. Advances in information
technology enable teachers to develop rapidly personalised learning
instruction materials and utilise real-time data for a feedback mechanism.
This learning paradigm promotes learner‟s choice in his education. It closes
the gaps in the individual learner, learning styles and provides the needed
support to succeed in his/her education. In this paradigm, teachers become
mentors or facilitators while learners have controls over what, how and
where they learn.

© 2018 The authors and IJLTER.ORG. All rights reserved.

 46

Ubiquitous Learning
With the emergence of ubiquitous learning (u-learning), access to education
by students have become flexible, and seamless. U-learning has been
described as a hybridisation of the best features of e-learning and m-learning
paradigms (Wong et al., 2015; Casey, 2005). Therefore, u-learning is a learning
paradigm that enables learning what is necessary at the right place and time
in the right way (Yahya et al., 2010). This learning paradigm was made
possible due to the availability of mobile, wireless communication and
sensing technologies (Hsu et al., 2016) which can support real-world contexts
learning (Yin et al., 2016). This learning paradigm has been described as the
context-aware ubiquitous learning or contextual mobile learning (Brown &
Mbati, 2015; Chen & Huang, 2012; Hwang et al., 2008).

Affective Computing
Affective computing can be referred to systems and device study and
development that can simulate, recognise, interpret, and process human
affects (Becker-Asano & Wachsmuth, 2010; Picard, 1997). Haijan et al. (2017)
proposes a study based on learner‟s interest and emotion recognition to
effectively construct affective education to optimise and enhance the teaching
effectiveness. Learner‟s voice, text and behaviour extraction was carried out
using Semantic Orientation-Pointwise Mutual information. Research on
recognising emotion has made great progress since affective computing was
proposed by Professor Picard in 1997 (Picard, 1997).

According to Kalyuga (2011), affective computing has been a fundamental

element in designing and developing adaptive learning systems. This is due
to the fact that learning process is complex (Bennet & Bennet, 2008) and
involves students‟ cognitive activities and emotions. Therefore, many
researchers suggested that effective and meaningful learning can be achieved
when students have positive cognitive as well as emotional states during
learning. From eLearning perspective, students‟ emotions will ensure the
effectiveness of a learning process (López-Pérez et al., 2011; Khan, Weippl, &
Tjoa, 2009).

Different methods for detection of emotion are the use of biophysical signals,
speech signals, facial expressions, machine learning techniques applied on
learners‟ preferences and behaviours in the systems, different text-based
method and virtual reality (Callaghan et al., 2009). The former methods
though give better accuracy, are quite expensive and cumbersome and are
less practical to implement in a large-scale e-learning environment, whereas,
the methods like machine learning, text-based methods and virtual reality are
less cumbersome and are more suitable in such scenario (Marczak et al.,
2016).

However, confidence (Baker et al., 2010), effort (Zimmerman, 2013),

independence (Baker, 2010) and confusion (Baker et al., 2010) have been used
to identify the pattern of behaviour of the learners‟ affective states. These four

© 2018 The authors and IJLTER.ORG. All rights reserved.

 47

affective states will be used in this paper to identify the pattern of behaviour
of the learners‟ affective states.

3. Methodology
This section discusses the system architecture and models used.
A. System Architecture
Figure 1 consists of the key components of the system architecture. These
include Learners location detection, learner interface, learners profile,
adaptive learning system, learning resources, learning activities, affect
recognition module, web server and evaluation module while the access
point (AP) wireless signal is used for detecting and calculating learners‟
location via mobile devices.

Learner interface and learners profiles consist of a registration, login

module and other personalised records which helps to extract and store
learner‟s preferences and also affect features. The adaptive learning
system helps to determine the cognitive characteristic, knowledge space
and preferences.

The learners use the learner interface to interact with the adaptive system
to get learning resources and participate in learning activities such as
forum discussion and content outline. The learners‟ profiles contain the
affect states and learning progress of individual learner. The adaptive
system also interacts with the learning resources which are stored in a
web server based on learner‟s profile. Learners‟ behaviours are captured
during the learning procedure. Affect recognition module help detect
what the learners‟ emotions are (Confidence, Effort, Independence and
Confusion) which are used to improve the system.
Detect Leaners Locations Personalisation

Learner Interface Adaptive Learning System Learning Resources

Learner’s
Learning Activities
Interact

Registration

Learner’s Profile Profiling

Interact

Activity Logs Feedback Feedback

Performance

Response Time Forum Discussion

Affect Recognition Module

Web Server

Affective status Affective status

Evaluation

Figure 1: Affect-Adaptive System Architecture

© 2018 The authors and IJLTER.ORG. All rights reserved.

 48

B. Affect-Adaptive learning models

Location
In order to detect learner‟ location, trilateration algorithm and
fingerprinting technique are employed to implement a received signal
strength (RSS) based location method (Lin et al., 2014; Setiya and Gaur,
2012). Their main aim is to estimate and get the learners location by
calculating the coordinates through geometry. In the geometry algorithm,
at least three of the anchor node and their estimated distance from the
node must be known (Ruiz et al., 2013; Doukhnitch et al., 2008). The
coordinates of the three nodes are represented by N1, N2, and N3 the
distances from these three anchor nodes to node N are dn1, dn2, dn3.
The coordinates of node N(x, y) are unknown, then formulae in equation
(1) is stated as follows (Sadoun & Al-Bayari, 2007):

dn1 = x − xn1 2+ y − yn1 2

dn2 = x − xn2 2+ y−y 2

n2
(1)
dn3 = x − xn3 2 + y − yn3 2

Using Equation (1), the coordinates of node N are calculated.

−1
x 2 xn1 − xn3 2 ya1 − yn3
y = X
2 xn2 − xn3 2 ya2 − yn3
2 2 2 2
xn1 − xn3 + yn1 − yn3 + d2n1 − d2n3
2 2 2 2
xn2 − xn3 + yn2 − yn3 + d2n2 − d2n3
(2)
According to equation 2 (Lin et al., 2014),
- Trilateration algorithm is used to calculate the target point
- The radius of the three Access Points (Aps) represent the distance
between the target and the corresponding APs
- Target point is the interception of the APs location.

Determining Learners’ Activities

Ant colony clustering algorithm is applied to this research work for
determining learners‟ activities, especially on learning resources. Ant
colony clustering algorithm is a clustering algorithm based on ant
cleaning behaviour, and the main idea is the process of transporting ants
(Hu et al., 2012). Ants start-off by being placed onto graph nodes. They
choose an edge to traverse from the home node depending on the
pheromone levels on each edge. Since the edge will not have any
pheromone yet, the ant will deposit pheromone equating to its features
when leaving a node that matches one or more features. The pheromone
deposition depends on the features which are present on the node and
also depends on the attractiveness of the features to the ant. Hence, with
each ant as an individual leaner, each learner determines the probability

© 2018 The authors and IJLTER.ORG. All rights reserved.

 49

of handling an activity according to the similarity between the learning

objects and the environment.
The similarity between different learning objects in the learning
environment can be defined as the arithmetic mean of the probability of
each attribute of the object as shown in equation (3):

n
f(Si ) = 1/n Pij (3)
j=1

Where n represents the number of attributes of the learning objects

The probability of an ant picking up can be defined as

1−e −cf (S i )
Pp = 1 − (4)
1+e −cf (S i )

While the probability of an ant dropping is defined as follows

1−e −cf (S i )
Pd = 1 − (5)
1+e −cf (S i )

c is a constant that is used to adjust convergence speed, once the c value is

different and the function of convergence speed is different.
Table 1 and Table 2 summarise the different patterns of behaviours to
recognize affective states and learning styles respectively. The patterns in
Table 1 are embedded for an individual dimension of the learning style
while the pattern in Table 2 incorporated for each affect-adaptive state to
achieve a personalised adaptive learning system. Based on the number of
occurrence of each pattern, the learners learning style, and specific
affective state are determined which are then used for calculating the
affective states and learning style preferences recognised.

Table 1: Patterns of behaviour to recognise learning styles

Active/Reflective Sensing/Intuitive Visual/verbal Sequential/global
contntvisit(-) contntvisit(-) questgraphics(+) outlinevisit(-)
contntstay(-) contntstay(-) questtext(-) outlinestay(-)
outlinestay(-) examplevisit(+) contntvisit(-) courseovviewvis
it(-)
forumcontntpost examplestay(+) contntstay(-) courseovviewsta
(+) y(-)
forumcontntpost slfassesvisit(+) forumcontntp navigationskip(-)
reply(+) ost(-)
forumassignmntp slfassessstay(+) forumcontntpost
ostrepl(+) repl(-)
slfassessvisit (+) exercsvisit(+)
slfassessstay(-) exercsstay(+)
exercsvisit(+) slfassessrevision(+)
exercsstay(+) assignmntrevision(+)
assignmntstay(+)

© 2018 The authors and IJLTER.ORG. All rights reserved.

 50

Table 2. Patterns of behaviour to recognise affective states

Confidence Effort Independence Confusion
contntvisit slfassessvisit contntvisit slfassessvisit
outlinevisit slfassessstay outlinevisit exercsvisit
examplevisit exercsvisit examplevisit examplevisit
exercsvisit exercsstay forumcontntvisit examplestay
forumassignmntpos
slfassessvisit forumcontntvisit forumcontntpost
t
forumcontntpostre forumcontntpostrep
forumcontntpost Assignmntrevision
pl l
forumcontntpostreply assignmnt revision assignmntrevision Contntstay
forumassignmntvisit slfassessvisit forumcontntvisit
forumcontntpost
forumcontntvisit slfassessstay
repl
forumcontntpost slfassessrevision

exercsvisit

forumassignmntpostrepl exercsstay

Graf et al. (2009) proposed an approach for calculating learning style from
behavioural pattern through hints calculation. Hints are represented by
four values that is 0-3, where 0 denotes weak indication towards the
respective learning style used for learning preferences. Learners‟
behaviour indicates average if the hint value is 2 (two), which means that
a specific hit is not provided. Also, the learners‟ behaviour is said to be in
disagreement with its learning style or affective state if the hint value is 1.
Hint value 0 indicates zero information on learners‟ behaviour.

To calculate learning styles and their respective affective states, equation

(6) is used
n
i=0 x i
; 0 ≤ xi ≤ 3 (6)
n
Where x represents a hint value for each patter, i represents pattern
number and i represent number of patterns giving information.

This measure is then normalised on a range from 0 to 1 for both learning

style and affective state. The value 1 represents a strong positive level and
the value 0 represents a strong negative level for the particular learning
style and affective state. In case no information is available for all patterns
of a learning style dimension or affective state, no conclusion can be
drawn.

4. Results And Discussion

One hundred and twenty–nine (129) learners took part in the online course.
The same set of students were enrolled in two different courses lasting for
about 5weeks. The two courses are course A (with affect) and course B (non-

© 2018 The authors and IJLTER.ORG. All rights reserved.

 51

affect). Data were collected through online self-report including measures of

emotional experience while relating with activities and resources online. The
obtained data were significant for this study because it contained timestamp
interaction activities recorded for each learner during learning. Data recorded
such as view discussion, post discussion, resources visit, content visit and
outline visit as are part of what made up the application logs. The log has
3552 activities for course A (with affect) and 2481 activities for course B (non-
affect).

Table 3 shows the breakdown of behavioural pattern capture in the system

log for course A and B were patterns such as examplevisit,
forumassignmntvisit and forumconentvisit has the highest percentage
records. Course A has a forumcontntpostreply of 308 for two forum post
while Course B has 209. Table 3 also shows that learners engaged more in
activities when their affective states are considered.

Table 3: Break down of Behavioural patterns

COURSE A COURSE B
Patterns of Behaviour Number Percentage(%) Number percentage(%)

Outlinevisit 33 1 21 1
Contentvisit 90 3 45 2
Contentstay 57 2 32 1
Assignmntrevision 24 1 2 0

Examplevisit 823 23 530 21

Exercsvisit 213 6 115 5
Exercsstay 322 9 206 8
Forumcontntpost 2 0 2 0

Forumcontntpostreply 308 9 212 9

Forumassignmntvisit 593 17 400 16

Forumcontntvisit 700 20 600 24

Slfassessvisit 211 6 190 8

Slfassessstay 133 4 121 5
Slfassessrevision 43 1 5 0

TOTAL 3552 100 2481 100

Response Timing.
The observations from this Table 4 are discussed as follows: The response
timing for course A with respect to Discussion 1 is that 64% of the learners
responded on the due date while. 4% responded after the due date while in
discussion 2, only 2% responded after the due date. This response suggests an
improvement of student understanding. In course B, there was a decrease in
the number of learners that submitted their responses by due date with only
32% and 41% for both discussion 1 and discussion 2 respectively.

© 2018 The authors and IJLTER.ORG. All rights reserved.

 52

Table 4: Timing of Learners Response Post

TIMING OF LEARNERS RESPONSE POST
COURSE A COURSE B
Timing Discussion 1 Discussion 2 Discussion 1 Discussion 2
Learners % of Learners % of Learners % of Learners % of Learners
Number Learners Number Learners Number Learners Number
Response after 41 32 36 28 1 1 15 12
a day
Response 83 64 67 52 41 32 53 41
posted on due
date
Response 5 4 3 2 75 58 61 47
posted after due
date

Quantity of Post
Individual participant is expected to reply to the forum post at least once in
the online forum discussion. There were about 308 valid responses for course
A as shown in Table 5 and about 212 valid responses for Course B. With both
Course A and Course B having an average of 2.39 posts per person and 1.64
posts per person responses respectively. Course A shows more intensity and
interest of learners to participate in the online discussion when compared to
course B. Course A benefitted immensely when learners‟ affective states are
considered.

Table 5: Number of responses posted by learners of course A and course B

Number
Course A
of Course A Course B Course B
Discussion 2
responses Discussion 1 Discussion 1 Discussion 2
posted

No. % No. % No. % No. %

0 6 5% 29 22 32 25 1 1%
1 110 85% 90 70 93 72 92 71%
2 10 8% 10 8 3 2 31 24%
3 3 2% 0 0 1 1 5 4%

Performance in Assessment Task

The Table 6 depicts the outcome of the performance of learners in a Course A
and Course B respectively. Grading the cohort of learners in course A
compared with performance Cohort of Course B reveals that for course A
about 8% scored below average while 70% performed above average for
Discussion 1 while about 64% performed above average for Discussion 2. In
Course B, 29% and 20% of the learners scored above average in discussion 1
and 2 respectively.

© 2018 The authors and IJLTER.ORG. All rights reserved.

 53

Table 6: Performance in Assessment Task

performance in assessment4
Course A Course B
Performance Discussion 1: Discussion 2: Discussion 1: Discussion 2:
(%)Learners (%)Learners (%)Learners (%)Learners
Number Number Number Number
below average 8% 5% 30 20
Average 22% 31% 41 60
satisfactory 40% 41% 27 13
distinction 30% 23% 2 7

From Table 7, the analysis was made according to how well the learners
participated online. Course A show more vibrant and aggressive participation
when compared to Course B.

Table 7: Break down of Behavioural patterns for Courses A and B

Cours
Confide Course Course Course Indepen Course Course Confusi Course Course
eB Effort
nce A Log. A Log B Log dence A Log B Log on A Log B Log
Log.

contntvi slfasse contntvis slfassess

90 45 211 190 90 45 211 45
sit ssvisit it visit
outlinev slfasse outlinevi exercsvi
33 21 133 121 33 21 213 21
isit ssstay sit sit
Exampl exercs examplev example
823 530 213 115 823 530 823 530
evisit visit isit visit
exercsvi exercs forumcon example
213 115 322 206 700 600 0 600
sit stay tntvisit stay
forum forumas
Slfasses forumcon
211 211 contnt 700 600 2 2 signmnt 2 2
svisit tntpost
visit post
forum
Forumc forumcon Assign
contnt
ontntpo 2 2 308 212 tntpostre 308 212 mntrevi 24 212
postre
st pl sion
pl
assign
forumco
mnt assignmn contntst
ntntpost 308 212 24 2 24 2 57 2
revisi trevision ay
reply
on
forumas
slfassessv forumco
signmnt 593 400 211 190 700 190
isit ntntvisit
visit
Forumc forumco
slfassesss
ontntvis 700 600 133 121 ntntpost 308 121
tay
it repl
slfassessr
43 5 5
evision
exercsvisi
213 115 115
t
exercssta
322 206 206
y
TOTAL 2973 2136 1911 1446 2902 2049 2338 2049

© 2018 The authors and IJLTER.ORG. All rights reserved.

 54

In order to determine whether the learners‟ affect-adaptive states would affect

learning performance or not, the analysis was done using statistical tool and
presented in Table 8 for both course A and Course B. Mean and Standard
Deviation (SD) were calculated according to the four(4) states (Confidence,
Effort, Independence and Confusion). Confidence state had the highest values
among all others with Mean=330.00 and SD=302.25 for course A and
Mean=237.33 and SD=223.36 for course B. This confidence value showed that
learning performance can be influenced either positively or negatively by
confidence state. The Confusion states also had a value with Mean=259.78
and SD=305.89 for course A and Mean=187.67 and SD=229.31 for course B.
The confusion values show that confused learners participate less on the
learning platform. The higher average mean and standard deviation (SD)
value for Course A (Mean=276.23, SD=272.27) over Course B (Mean=200.58,
SD=210.51) shows an improvement in learning performance. Also, a t-Test
analysis was carried out to test the statistical significance difference. The test
confirms that the difference was significant and the t-Test result suggested
learners‟ performance was dependent on their affect-adaptive states.

Table 8. Statistical details of learners’ Affect-Adaptive states

Affect states Course A Course B t TEST

Mean SD Mean SD T Stat p d
Confidence 330.33 302.25 237.33 223.36 2.858 0.021 8
Effort 273 214.09 206.57 188.26 3.857 0.008 7
Independence 241.83 266.85 170.75 201.11 3.039 0.011 11
Confusion 259.78 305.89 187.67 229.31 2.342 0.047 8
AVERAGE 276.23 272.27 200.58 210.51
VALUE

5. Conclusion
This research work was conducted to identify learners‟ affect-adaptive states and
see how learning performances can be improved using these states in a
personalised ubiquitous learning system. The use of ubiquitous technology
makes accessibility to education resources available at anywhere and anytime.
Considering adaptability, personalisation and ubiquity, this research paper
investigates learners’ affective states such as confidence, effort, independence
and confusion when learning. Trilateration and fingerprint techniques are both
used to determine the learners’ locations. The learners’ preference is determined
using Felder Silverman learning style dimensions based on learners’ activities in
online learning process. The ant colony clustering algorithm is used to determine
learners’ activities. The four affective states are determined from the learners’
forum threads, content activities, assessment, timing, and discussion through
learners’ engagement. The result of the study suggests that performance of
learners improves when learners‟ affective states are considered during learning.

© 2018 The authors and IJLTER.ORG. All rights reserved.

 55

6. Acknowledgement
The research is sponsored by Tertiary Education Trust Fund (TETFUND) in the
context of “Personalised Adaptive Ubiquitous Learning System” under the
National Research and Innovation Grant.

7. References
Adewale O. S. (2007), University Digital Libraries: adaptive, personalised e-learning
system, Yerimah Prints, ISBN: 978-38410-2-5
Baker, R. S. J. D. (2010). Data mining for education. International encyclopedia of education,
7(3), 112-118.
Baker, R. S., D'Mello, S. K., Rodrigo, M. M. T., & Graesser, A. C. (2010). Better to be
frustrated than bored: The incidence, persistence, and impact of learners‟
cognitive–affective states during interactions with three different computer-
based learning environments. International Journal of Human-Computer Studies,
68(4), 223-241.
Becker-Asano, C., & Wachsmuth, I. (2010). Affective computing with primary and
secondary emotions in a virtual human. Autonomous Agents and Multi-Agent
Systems, 20(1), 32.
Behera, S. K. (2013). E-and M-Learning: A comparative study. International Journal on New
Trends in Education and Their Implications, 4(3), 65-78.
Beldagli, B. & Adiguzel, T. (2010). “Illustrating an ideal adaptive e-learning: a conceptual
framework,” in Proceedings of the 2nd World Conference on Educational
Sciences (WCES '10), pp. 5755–5761.
Bennet, A., & Bennet, D. (2008). E-Learning as energetic learning. VINE, 38(2), 206-220.
Besterfield-Sacre, M., Amaya, N.Y., Shuman, L. J., Atman, C.J., & Porter, R. (1998).
Understanding student confidence as it relates to first year achievement”,
Frontiers in Education Conference, November, pp 258-263
Brown, T. H., & Mbati, L. S. (2015). Mobile learning: Moving past the myths and
embracing the opportunities. The international review of research in open and
distributed learning, 16(2).
Callaghan, V., Shen, L., Gardner, M., Shen, R., & Wang, M. (2009). A mixed reality
approach to hybrid learning in mixed culture environments. Handbook of Research
on Hybrid Learning Models: Advanced Tools, Technologies, and Applications:
Advanced Tools, Technologies, and Applications, 260.
Casey, D(2005). In E-Learn 2005–World Conference on E-Learning in Corporate,
Government, Healthcare, and Higher Education; Richards, G., Ed.; Association
for the Advancement of Computing in Education (AACE): Vancouver, BC,
Canada, 2005; pp. 2864–2871.
D‟Mello, S.K., Lehman, B., Pekrun, R., & Graesser, A.C (2014). Confusion can be
beneficial for learning. Learn. Instr. 29(1), 153–170.
Doukhnitch, E., Salamah, M., & Ozen, E (2008). An efficient approach for trilateration in
3D positioning, Computer Communications, 31(17), 4124–4129.
Entwistle, N., & Ramsden, P. (2015). Understanding student learning (Routledge revivals).
Routledge.
Farman, A.K., Sabine, G., Edgar, R. W., & A Min T. (2010). Identifying and incorporating
affective states and learning styles in web-based learning management systems,
Interaction Design and Architecture(s) Journal, vol. 9, no. 10, pp. 85–103.
Fasihuddin, H., Skinner, G., & Athauda, R. (2017). Towards adaptive open learning
environments: Evaluating the precision of identifying learning styles by tracking
learners‟ behaviours. Education and Information Technologies, 22(3), 807-825.

© 2018 The authors and IJLTER.ORG. All rights reserved.

 56

Feidakis, M., Caballé, S., Daradoumis, T., Jiménez, D. G., & Conesa, J. (2014). Providing
emotion awareness and affective feedback to virtualised collaborative learning
scenarios. International Journal of Continuing Engineering Education and Life Long
Learning 6, 24(2), 141-167.
Felder, R. M., & Silverman, L. K. (1988), Learning and teaching styles in engineering
education, Engineering Education, 78(7): 674–681
Hsu, T.Y., Chiou, C.K., Tseng, J.C.R., & Hwang, G.J. (2016). Development and Evaluation
of an Active Learning Support System for Context-Aware Ubiquitous Learning.
IEEE Trans. Learn. Technol. 2016, 9, 37–45.
Hu, X.H., Mu, T., Dai, W.H., Hu, H.Z., & Dai, G.H(2012). Analysis of browsing behaviors
with ant colony clustering algorithm. Journal of Computers 7(12), 3096–3102.
Hwang, G.-J., Tsai, C.-C., & Yang, S.J.H. (2008). Criteria, Strategies and Research Issues
of Context-Aware Ubiquitous Learning. J. Educ. Technol. Soc., 11, 81–91.
Jones, V., & Jo, J. H. (2004). Ubiquitous learning environment: An adaptive teaching
system using ubiquitous technology. In R. Atkinson, C. McBeath, D. Jonas-
Dwyer, & R. Phillips (Eds.), Proceedings of the 21st ASCILITE (Australasian
Society for Computers in Learning in Tertiary Education) Conference (pp.
468e474). Perth, Australia: ASCILITE.
Kalyuga, S. (2011). Cognitive load theory: Implications for affective computing. Paper
presented at the FLAIRS Conference.
Keltner, D., & Shiota, M.N. (2003), New displays and new emotions: A commentary on
Rozin and Cohen (2003) Emotion, 3(1): 86-91
Khan, F. A., Graf, S., Weippl, E. R., & Tjoa, A. M. (2010). Identifying and Incorporating
Affective States and Learning Styles in Web-based Learning Management
Systems. IxD&A, 9, 85-103.
Khan, F. A., Weippl, E. R., & Tjoa, A. M. (2009). Integrated Approach for the detection of
learning styles and affective states. In World Conference on Educational
Multimedia, Hypermedia and Telecommunications, (1), 753-761.
Kim, S. B. (2006). Cyber communication & ubiquitous language learning: Feasibility &
reality with reference to survey research. Multimedia Assisted Language Learning,
7(3), 9-45.
Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance during
instruction does not work: An analysis of the failure of constructivist, discovery,
problem-based, experiential, and inquiry-based teaching. Educational
psychologist, 41(2), 75-86.
Kort, B., Reilly, R., Picard, R.W. (2001). An affective model of the interplay between
emotions and learning. In: Proceedings of the IEEE International Conference on
Advanced Learning Technologies (ICALT „01). IEEE Computer Society, pp. 43–
46.
Lin, P., Li, Q., Fan, Q., Gao, X., & Hu, S. (2014). A real-time location-based services
system using WiFi fingerprinting algorithm for safety risk assessment of
workers in tunnels. Mathematical Problems in Engineering, 2014.
Liyanage, M. P. P., KS, L. G., & Hirakawa, M. (2016). Detecting learning styles in learning
management systems using data mining. Journal of Information Processing, 24(4),
740-749.
Lopatovska, I., & Arapakis, I. (2011). Theories, methods and current research on
emotions in library and information science, information retrieval and human–
computer interaction. Information Processing & Management, 47(4), 575-592.
López-Pérez, M. V., Pérez-López, M. C., & Rodríguez-Ariza, L. (2011). Blended learning
in higher education: Students‟ perceptions and their relation to outcomes.
Computers & education, 56(3), 818-826.

© 2018 The authors and IJLTER.ORG. All rights reserved.

 57

Marczak, M., Krajka, J., & Malec, W. (2016). Web-based assessment and language
teachers-from Moodle to WebClass. International Journal of Continuing
Engineering Education and Life Long Learning, 26(1), 44-59.
Moore, M. G., & Kearsley, G. (2011). Distance education: A systems view of online learning.
Cengage Learning.
Paramythis, A. & Loidl-Reisinger, S. (2004). Adaptive learning environments and e-
learning standards. Electronic Journal of eLearning, 2(1), March.
http://www.ejel.org/volume-2/vol2-issue1/issue1-art11.htm
Pintrich, P.R. & DeGroot, E.V. (1990), Motivational and self-regulated learning
components of classroom academic performance, Journal of Educational
Psychology, 82: pp. 33-40.
Premlatha, K. R., Dharani, B., & Geetha, T. V. (2016). Dynamic learner profiling and
automatic learner classification for adaptive e-learning environment. Interactive
Learning Environments, 24(6), 1054-1075.
Qu, L., Wang, N., Johnson & Lewis, W. (2005). Using Learner Focus of Attention to
Detect Learner Motivation Factors, In Ardissono, L., Brna, P., Mitrovic, A.: User
Modelling, pp 70-73.
Rozin, P. & Cohen, A.B. (2003a). Confusion infusions, suggestives, correctives, and other,
medicines, Emotion, 3: 92-96
Sadoun, B. & Al-Bayari, O. (2007). Location based services using geographical
information systems, Computer Communications, 30(16), .3154–3160.
Sangineto, E., Capuano, N., Gaeta, M., & Micarelli, A. (2008), Adaptive course generation
through learning styles representation, Journal of Universal Access in the
Information Society 7(1), 1–23.
Setiya, R., & Gaur A. (2012). Location fingerprinting of mobile terminals by using Wi-Fi
device, International Journal of Advanced Research in Computer Engineering &
Technology, 1(4), 311–314.
Siadaty, M. & Taghiyareh, F. (2007). PALS2: Pedagogically Adaptive Learning System
based on Learning Styles, Seventh IEEE International Conference on Advanced
Learning Technologies (ICALT), pp.616-618.
Sung, Y. T., Chang, K. E., & Liu, T. C. (2016). The effects of integrating mobile devices
with teaching and learning on students' learning performance: A meta-analysis
and research synthesis. Computers & Education, 94, 252-275.
Tortorella, R. A., & Graf, S. (2017). Considering learning styles and context-awareness for
mobile adaptive learning. Education and Information Technologies, 22(1), 297-315.
Tummala, H., and Jones, J. (2005). Developing spatially-aware content management
systems for dynamic, location-specific information in mobile environments. The
3rd ACM International Workshop on Wireless Mobile Applications and Services
on WLAN Hotspots (pp. 14–22).
Wang, H. Y., Liu, G. Z., & Hwang, G. J. (2017). Integrating socio‐ cultural contexts and
location‐ based systems for ubiquitous language learning in museums: A state
of the art review of 2009–2014. British Journal of Educational Technology, 48(2), 653-
671.
Wanner, T., & Palmer, E. (2015). Personalising learning: Exploring student and teacher
perceptions about flexible learning and assessment in a flipped university
course. Computers & Education, 88, 354-369.
Wong, L. H., Milrad, M., & Specht, M. (Eds.). (2015). Seamless learning in the age of mobile
connectivity. Singapore: Springer.
Yahya, S., Ahmad, E.A., & Jalil, K.A. (2010). The definition and characteristics of
ubiquitous learning: A discussion. Int. J. Educ. Dev. Using Inf. Commun.
Technol., 6, 117–127.

© 2018 The authors and IJLTER.ORG. All rights reserved.

 58

Yin, P.-Y., Chuang, K.-H., & Hwang, G.-J. (2016). Developing a context-aware ubiquitous
learning system based on a hyper-heuristic approach by taking real-world
constraints into account. Univers. Access Inf. Soc., 15, 315–328.
Yuan, L., Powell, S., & CETIS, J. (2013). MOOCs and open education: Implications for
higher education.
Zimmerman, B. J. (2013). Theories of self-regulated learning and academic achievement:
An overview and analysis. In Self-regulated learning and academic achievement (pp.
10-45). Routledge.

© 2018 The authors and IJLTER.ORG. All rights reserved.

A. Identitas

Judul : Affect-Adaptive Activities in a Personalised

Ubiquitous Learning System
Tahun : 2018
Penulis : O. S. Adewale, O. C. Agbonifo, E. O. Ibam, A. I.
Makinde, O. K. Boyinbode, B. A. Ojokoh and O.
Olabode
Nama jurnal /Terbitan No : International Journal of Learning, Teaching and
Educational Research Vol. 17, No. 7, pp. 43-58
Data indeks : Quartile 3

B. Pembahasan dan Analisa

Fase perkembangan teknologi informasi telah mengubah sistem Pendidikan, Perkembangan

ini memberikan akses komunikasi global dan sumber daya yang tersedia bagi siswa saat ini di semua
tingkat sekolah, Gaya belajar memungkinkan untuk menetapkan cara belajar yang disukai. Setiap
kali seorang pembelajar dengan gaya belajar yang unik tidak didukung oleh lingkungan pengajaran
apa pun, hal itu membuat pembelajaran menjadi sulit bagi pembelajar tersebut Misalnya, pelajar
belajar lebih banyak ketika instruksi yang disajikan sesuai dengan gaya belajar mereka Metode
terbaik untuk mendapatkan gaya belajar yang didasarkan pada materi instruksional di kelas besar
adalah melalui lingkungan e-learning yang adaptif karena tidak memiliki keterbatasan umum yang
dimiliki skenario tatap muka yang berfokus pada siswa individu karena kurangnya sumber daya dan
waktu yang dibutuhkan Penerapan komputasi di mana-mana dalam pendidikan yang dikenal
sebagai
pembelajaran di mana-mana (u-learning) menandai langkah maju yang besar dalam penyampaian
materi instruksi pendidikan. u-learning telah merevolusi pendidikan dan mengatasi beberapa kendala
pembelajaran tradisional. Namun, pembelajaran menggunakan konten pembelajaran yang sama
akan menyebabkan ketidakcocokan kognitif. Integrasi pembelajaran adaptif dan komputasi di
mana-mana dapat menawarkan inovasi besar dalam penyampaian pembelajaran, memberikan akses
ke personalisasi dan penyesuaian dengan kebutuhan siswa (Sung et al., 2016; Kim, 2006).
 Lebih jauh lagi, dalam lingkungan belajar tradisional, seorang guru merasa sulit untuk mengajar
semua peserta didik secara individual karena jumlah peserta didik yang banyak. Namun, guru dapat
memantau dan menarik kesimpulan dari perilaku belajar siswa yang dapat disebut sebagai kondisi
afektif siswa. Sistem pembelajaran yang mempertimbangkan keadaan afektif peserta didik akan
meningkatkan pengalaman belajar peserta didik (Baker et al., 2010).
Beberapa parameter seperti kemarahan, ketakutan, kegembiraan, keterkejutan, rasa jijik,
motivasi, minat, kecenderungan, kepercayaan diri,kebingungan, dan usaha telah digunakan untuk
menggambarkan keadaan afektif peserta didik (Khan et al., 2010; Qu et al., 2005; Lopatovska &
Arapakis, 2011). Afek telah terbukti mempengaruhi proses pembelajaran (Kort et al. 2001; D
"Mello et al. 2014; Baker et al. 2010). Keadaan afektif dapat diklasifikasikan menjadi dua, yaitu
keadaan afektif positif dan keadaan afektif negatif. Keadaan afektif positif seperti kepercayaan diri,
kepuasan dan sebagainya berkontribusi terhadap pembelajaran sedangkan keadaan afektif negatif
menghambat pembelajaran. Oleh karena itu, makalah ini menyelidiki pengaruh adaptivitas,
personalisasi, ubiquity, dan keadaan afektif pada pembelajaran siswa.
Emosi dalam pembelajaran karena pengaruhnya terhadap kinerja pelajar sehingga memberikan
alasan mengapa keadaan afektif pelajar sangat penting untuk pembelajaran. Dalam hal ini, penelitian
ini bertujuan untuk mengidentifikasi aktivitas afektif-adaptif yang akan meningkatkan pembelajaran
dalam sistem pembelajaran yang dipersonalisasi di manamana dengan mendeteksi gaya belajar dan
keadaan afektif melalui beberapa pola perilaku. Dalam penelitian ini, preferensi peserta didik
ditentukan, keadaan afektif peserta didik seperti kepercayaan diri, usaha, kemandirian dan
kebingungan diselidiki. Algoritma pengelompokan koloni semut digunakan untuk menentukan
aktivitas peserta didik. Keempat kondisi afektif ditentukan dari thread forum peserta didik, aktivitas
konten, penilaian, waktu, dan diskusi melalui keterlibatan peserta didik. Hasil dari penelitian ini
pada dua mata kuliah yang terpisah, Mata Kuliah A (Pengaruh) dan Mata Kuliah B (Non-Pengaruh)
menunjukkan bahwa keempat keadaan afektif mempengaruhi pembelajaran

C. Rangkuman

munculnya pembelajaran di mana-mana (u-learning), akses ke pendidikan oleh siswa menjadi

fleksibel dan tanpa batas. Definisi u-learning dapat diklasifikasikan secara luas ke dalam tiga
kategori: teknologi yang diadopsi, objek pendidikan, dan proses pembelajaran (Peña-Ayala &
Cárdenas, 2016). Dalam definisi yang dipengaruhi oleh klasifikasi lingkungan belajar, u-learning
digambarkan sebagai kombinasi dari paradigma e-learning dan m-learning (Behera,2013).

Seorang pelajar yang gaya belajarnya tidak didukung oleh lingkungan belajar tertentu pasti akan
menghadapi kesulitan selama pembelajaran (Entwistle & Ramsden, 2015; Felder dan Silverman,
1988).

Felder dan Silverman Learning Style Model (FSLSM) adalah salah satu model gaya belajar yang
digunakan dalam penelitian terkait gaya belajar untuk sistem pembelajaran berbasis teknologi
informasi. FSLSM telah terbukti paling sesuai untuk courseware hypermedia (Tortorella & Graf,
2017; Liyanage et al., 2016; Fasihuddin et al., 2017). FSLSM membedakan antara empat dimensi
yang berbeda. Dimensidimensi ini adalah dimensi aktif, reflektif, penginderaan, intuitif, visual,
verbal, sekuensial dan global. Kombinasi dari dimensi-dimensi ini membentuk preferensi belajar
individu peserta didik.. Kemajuan teknologi informasi telah banyak mengubah praktik
pembelajaran dalam beberapa tahun terakhir. Namun, platform pembelajaran seperti Moodle,
Blackboard, dan lain-lain tidak mendukung adaptivitas penuh yang didasarkan pada gaya belajar
dan keadaan afektif (Jones & Jo, 2004).

Pembelajaran adaptif dapat dirujuk pada konsep metodologi pembelajaran yang disesuaikan
dengan gaya belajar individu peserta didik. Beberapa elemen pembelajaran adaptif telah ditunjukkan
meliputi: aktivitas pemantauan, menafsirkan hasil, memahami persyaratan dan preferensi, dan
memfasilitasi proses pembelajaran dengan informasi yang diperoleh (Zimmerman, 2013;
Paramythis & Loidl-Reisinger, 2004).

Jenis sistem pembelajaran ini menawarkan keuntungan besar dalam memberikan pengetahuan
yang spesifik dan personal kepada peserta didik selama pembelajaran. Penggunaan strategi adaptif
dapat ditanggalkan kembali sejak awal abad kedua puluh ketika pendidik mulai mencari cara lain
untuk meningkatkan hasil belajar pemahaman dan pembelajaran siswa (Moore & Kearsley, 2011;
Zimmerman, 2013). Pengembangan system semacam itu pasti akan memberikan atau mendukung
pengalaman belajar yang efektif untuk jenis pembelajar yang berbeda di seluruh pandangan domain
pengetahuan yang lebih luas. Oleh karena itu, sistem pembelajaran adaptif harus menampilkan
preferensi, minat, konten pembelajaran, dan perilaku penelusuran peserta didik untuk menyediakan
layanan pembelajaran yang adaptif dan personal (Premlatha, 2016; Wang et al., 2017; Beldagli &
Adiguzel, 2010).
 Dengan munculnya pembelajaran di mana-mana (u-learning), akses ke pendidikan oleh siswa
menjadi fleksibel dan tanpa batas. U-learning telah digambarkan sebagai hibridisasi fitur terbaik
dari paradigma e-learning dan m-learning (Wong et al., 2015; Casey, 2005). m-learning
(Wong et al., 2015; Casey, 2005). Oleh karena itu, u-learning adalah paradigma pembelajaran
yang memungkinkan mempelajari apa yang diperlukan di tempat dan waktu yang tepat dengan
cara yang tepat (Yahya et al., 2010). Paradigma pembelajaran ini dimungkinkan karena
tersedianya teknologi komunikasi dan penginderaan nirkabel yang mobile (Hsu et al.,
2016) yang dapat mendukung pembelajaran konteks dunia nyata (Yin et al., 2016). Paradigma
pembelajaran ini telah dideskripsikan sebagai context- aware ubiquitous learning atau
contextual mobile learning (Brown & Mbati, 2015; Chen & Huang, 2012; Hwang et al., 2008).

Gambar 1. Arsitektur Sistem Afektif-Adaptif

D. Rekomendasi (rekomendasi untuk Pendidikan Indonesia)

Penelitian ini dapat di jadikan rujukan agar guru dapat mengidentifikasi keadaan afektif-adaptif
pembelajar dan melihat bagaimana kinerja pembelajaran dapat ditingkatkan dengan menggunakan
sistem pembelajaran di mana-mana yang dipersonalisasi. Selain itu Penggunaan teknologi di mana-
mana membuat aksesibilitas ke sumber daya pendidikan tersedia di mana saja dan kapan saja, serta
guru dapat menyelidiki keadaan afektif pembelajar seperti kepercayaan diri, usaha, kemandirian dan
kebingungan saat belajar. Penggunaan teknologi juga seperti Teknik trilaterasi dan sidik jari digunakan
untuk menentukan lokasi pembelajar.