Adaptive Learning Systems based on ILOs of Courses
MARWA HUSSIEN MOHAMED1, LAMIAA F. IBRAHIM2,3, KHALED ELMENSHAWY1,
HAITHAM RIZK FADLALLAH1
1Information System Department, October 6 University, Cairo, EGYPT
2Computer Science Department, October 6 University, Cairo, EGYPT
3Computer Science Department, Cairo University, Cairo, EGYPT
Abstract - Nowadays, the use of e-learning techniques and methods is a very important challenge due to the
importance of digital transformation to all countries. Firstly, the spread of the COVID-19 virus all over the world.
Secondly, all students need to study their courses remotely from home to reduce the communication with others to
save their life. All teachers need to engage their students effectively to study an online course, get more knowledge
and high results at the end of these courses. Data mining is the best tool used to find a hidden pattern. We used an
educational data mining tool to help teachers find the pros and cons of using an e-learning course with their
students. We need to classify students on these online courses according to their ability to understand materials and
quizzes, or assessment methods of the course, by making adaptive e-learning courses. In this paper, we will show
the importance of using adaptive e-learning courses and the challenges faced by authors to build these systems, and
we will list the different methods used with adaptive learning like gamification, brain-hex models, facial emotions,
and we will also list a survey about other authors' techniques and methods used to find the student's learner style.
We build a new proposed model of ILOs(Intended Learning Outcomes) adaptive learning with the emotion-based
system to let the system find the student's learning style and build the material according to their skills and
knowledge outcomes from the course and engage the use of facial emotion while taking the quiz to predict the
student's results and the topics he/she needs to study more via our system to achieve high grades and knowledge.
Our system finds that the visual students have the highest grades with 75%, followed by kinesthetic with 70% and
the lowest grades in auditory with 50%.
Keywords- e-learning; education data mining; classification; gamification courses, Brain-hex model, facial
emotions, Intended Learning Outcomes (ILOs).
Received: September 18, 2022. Revised: November 4, 2022. Accepted: December 7, 2022. Published: January 2, 2023
1 Introduction
Data mining, [1], is a process of analyzing large
students’ data and finding hidden patterns and
knowledge that can be utilized to help in learning
systems. It uses different classification techniques on
the student’s data and tracks their behavior while
using the adaptive learning system.
Educational data mining is an important discipline
to help the improvement of education systems by
predicting students’ behavior towards the online
system. Today, educational institutions collect and
archive massive amounts of data like registration data
for every semester; attendance; total classwork;
section grades; and final exam results to calculate the
final GPA.
There are a lot of data mining techniques, [2],
used in adaptive learning, like Naive Bayes
Algorithm, Linear and Logistic Regression, machine
learning techniques such as supervised learning
(Support Vector Machine (SVM) and decision tree
(DT)), and unsupervised learning (Clustering in
Fuzzy Logic), artificial neural network (ANN),
association rule mining, etc.
Teachers and instructors control their students in
the traditional classroom by tracking their attitudes
and facial emotions or reactions to the material. This
allows the teacher to know if the students understand
the material, he discusses with them. Also, traditional
e-learning systems were uploading PowerPoint
material to the students without any tracking of their
learning styles. The number of students registered in
courses is increasing every day, [3], [4]. We need to
make predictions for students' behavior online and
analyze this huge amount of information.
Today, the E-learning system, [5], increases the
engagement of students by building an adaptive
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
1
Volume 18, 2023
learning system to track the student’s behavior and
learning progress via new systems like the live
classroom and Moodle websites in universities or
schools. It will be more helpful to get the best final
exam course grades at the end of the semester.
An adaptive e-learning system, [6], [7], gives the
students opportunities to select which type of
learning material they need to study according to
their profile, interests, and previous knowledge of
this course. The adaptive learning systems change the
traditional methods like uploading presentation
materials only to all students and let instructors track
students' activity through the log data.
An adaptive e-learning system, [6], [8], [9], builds
a classifier model to group students according to the
learner style like the VAK model (Visual learners—
Auditory learners—Kinesthetic learners) and Felder
and Silverman Learning Style Model (FLSM), etc.
This paper is organized as follows In Section 2 we
will discuss the data mining techniques used in
education. In Section 3 we will show the importance
of adaptive e-learning and the challenges facing the
authors in building their models, and the more
important classifier brain hex model used by the
authors to find their learner style. In Section 4 we
will list how authors use games to build adaptive
gamification learning systems, Section 5 lists a lot of
related work papers about adaptive learning and
techniques used and lists the number of students
engaged in every experiment; Section 6 describes
how authors used facial emotions with adaptive
learning systems; Section 7 describes our new
proposed system ILOs adaptive learning with the
emotion-based system, Section 8 discuss our
experimental results and finally, the conclusion and
future work.
2 Data Mining In Education
Data mining in education comes from different
sources like machine learning, information
visualization psychometrics, and other areas of
computational modeling and statistics
Education mining techniques are divided into
three parts: statistics, visualization, and web mining.
We must classify students’ data and e-learning
material using the following steps, [1]:
1) We need to identify the relationship between what
we will store in the database and the information we
collect from the students. To find the relationships between
students, we will use different algorithms to predict the
learner style by using classification, regression, and
density estimation.
2) Clustering: we will create a group of students
with the same behavior in the system based on their
activity and learner style.
3) Relationship mining: to find the relationships
between students, we predict the student's learning styles
to improve their knowledge in this course and make
changes in the teaching process methods and interactive
materials.We can use one of the following techniques:
association rule mining , Correlation mining , Causal DM
and Sequential pattern mining.
4) Distillation of data for human judgment: it aims
to make educational data understandable and help the
human brain find new knowledge by presenting the data in
different ways and using a lot of visualization methods.
5) Discovery with models.
Instructors are responsible for creating the
presentations, pdf files, assignments, videos, quizzes,
etc. that discuss the material to all students with
different learner styles. Figure 1 shows the steps
between the lectures and students while they interact
and communicate with the classrooms and e-learning
systems. The use of data mining techniques with
education will improve the educational process and
the engagement of the students and adaptive systems
for different types of learners will get a good
understanding of the course.
Fig. 1. Data mining techniques for students and
lecturers, [1].
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
2
Volume 18, 2023
We have two types of data used in adaptive
educational systems, [6]:
● Structured data: personal data, learning
management system, and performance evaluation
data. Figure 2 shows the user data information.
● Unstructured data: Web data, social networking
data, audio files, and learning video data.
Fig. 2. User data information (structured data).
3 Adaptive Learning
Adaptive learning refers to building a model aiming
to extract useful information from an online system
adapted based on the student’s activity on it.
3.1 Adaptive Learning Importance in
Education
The important question we must ask before we start
building an adaptive e-learning system:
1) What? kind of data collected from the system
while the students start learning on it.
2) Who? We need to know the ages of the students.
We will build the system for them and know some of their
characteristics.
3) How? Adaptive systems analyze this collected
data and the method used to classify students according to
their interests.
4) A success target. We will achieve success when
our students achieve high scores via the new system
compared with traditional classrooms.
The e-learning data sources are:
● The log data records and activity describe the
student's interaction with the system and training
materials.
● The performance records measure the student's
results in passing the evaluation tests.
● The student's profile information will be used to
classify students and the way they learn from the
online system.
There are two analytics types for adaptive e-learning:
● Descriptive: it’s provided past context and enables
decisions that may influence next learning
processes.
● Predictive: Make educated guesses about aspects
and variables that may have an impact on current
learning processes. That will enable teachers to
take good actions about the next content that will
appear to students.
The adaptive learning system process as in Figure
3 starts by using the machine learning part to detect
the learning path based on the learner's requirements
and needs. The learning path is based on the previous
user’s interaction with the system and which ones are
similar enough in characteristics to recommend the
same learning style.
Fig 3: Adaptive E-learning System Architecture.
The Resource Description Framework is used
with machine learning to extract the database about
the learner's profile data and his log data on the
system during the learning process and the time
online on the system.
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
3
Volume 18, 2023
While we track his behavior on the system to
adapt it and recommend a specific material that
outcome from the machine learning interferences
with the system.
In the adapted system, we need to store static and
dynamic information about the student’s profile and
collect the learning and assessment process on the
system.
The final target we need to achieve from the
adaptive e-learning systems in the education process
is that we need to measure student performance after
the end of the semester by:
● Reducing student dropouts
● Improving students' learning and
understanding
● find which content is relevant for a given user
● Improving training and learning materials
3.2 Data Mining Techniques used to Solve
Adaptive E-Learning Challenges
We will list some of the adaptive e-learning
challenges and data mining techniques used to solve
these challenges:
1) The coursework and exam-based assessment gap
problem. They used Random Forest and Naive Bayes
classifiers, [10].
2) How to predict student performance while using the
system to learn They used supervised machine
learning, unsupervised machine learning, SVM, and
linear regression, [11].
3) How to help students in selecting their courses? They
make use of the J48 and K-Means algorithms, [12].
4) How to group students according to their activity on
the system. They use the fuzzy technique. Naive Bayes
Classifier, Tree C4.5 Algorithm, [13], Linear Fuzzy
Real Logistic Regression.
5) Predict the rate of graduation for the students. They
use Decision Trees, J48, and Random Trees, [13].
6) predict the impact of the student's performance on the
system and the measure results of their final
assessment. They used Random Forest, Decision Tree,
Naive Bayes, and Logistic Regression, [14].
7) How the institutions select better teaching and
learning methods, like making a suitable timetable
and teaching courses. They use K-means, Apriori
Algorithm, [15], [16].
8) How to predict which students are likely to fail a
course by tracking their assessment grades early in
the system and increasing their learning time to
reduce failure records. They use SVM (support vector
machine classifier), [17].
9) Finally, how to build a system and find the outlier
students that match the admission selection criteria.
They use a data mining admission model(DMAM)
using Rule Mining, [18].
3.3 BrainHex Model Classifier
BrainHex, [19] ,is one of the most popular surveys
used today with e-learning, adaptive learning, and
gamification websites to detect the users’ styles and
then build a system adapted to their preferences and
personality.
BrainHex is the result of many years of studying
neurobiological research papers and trying their
findings to the reality of game design and player
satisfaction modeling, as well as several previous
surveys into gaming such as the first demographic
game design model DGD1 survey (which resulted in
the DGD1 model) and the DGD2 survey (which
directly affected the development of BrainHex).
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
4
Volume 18, 2023
Table 1. Description of brainhex model, [19].
BrainsHex’s
Model
Description
Achiever
Goal-oriented and motivated by
completion. They like to collect and
complete everything they can find.
They prefer to carry out a series of
tasks within their reach, distinct
from Conquerors, who prefer to
overcome difficult obstacles.
Conqueror
Enjoy struggling against strong
opponents until they achieve victory.
They channel their anger to achieve
victory.
Daredevil
Motivated by excitement and risk-
taking by playing on the edge. They
enjoy rushing around at high speed
while still being in control of the
experience.
Mastermind
Solving puzzles, devising strategies,
and making the most efficient
decisions. They feel rewarded for
making well-thought decisions.
Seeker
Exploring the game world and
enjoying moments of wonder. This
motivation comes from the parts of
the brain that process sensory
information and memory association
Socialiser
Interacting with other people, talking
to them, helping them, or just
hanging around. They are trusting
and their behavior connects to their
social center in the brain.
4 Gamification in E-Learning
Gamification methods and ideas are used in learning
to motivate students in their learning process to
complete the course, like finishing a game level with
badges, leaderboard, and stars.
4.1 Gaming Categories in Learning Systems
The four most popular types of gaming used in e-
learning, [20]:
1) Game-based learning acts like actual games
added to the teaching methods in the classroom
and the use of video games to gain high attention
and motivation of the students during the
learning process.
2) A Serious game is an adaptive gaming method
used in schools to build a link between
technology and pedagogy. It appears to the
students like an ordinary game.
List of serious games types, [21]:
● Teaching game: it is used in a full game environment
to teach a concept of learning materials.
● Simulator game: it offers a virtual version of a stable
practice and testing of an object from the real world.
● Meaningful game: it conveys the meaning with a
meaningful message.
● Purposeful game: it needs to increasing users'
activities and evaluate or measure users’ ability to
creates a direct real-world.
3) Gamification, [20], in education is a serious
approach to accelerate the curve of the learning
experience, teach complex subjects, and systems
of thought”
4) Simulation: it’s used to let students try the
simulator system on the experiments and has
many various input variables, observe the
outputs, and record the results just like in a real
laboratory. It allows students to try more
practices safely and gain high benefits from this
learning method.
4.2 Types of Gamers
Much research is done on how to detect the user’s
player types in the gamified learning system and
categorize them based on his interaction with the
system environment. The individuals' player, [22],
[23], characteristics depend on context and
environment. The important criteria that must be
taken into consideration are the relationships between
the users and their preferences.
The work, [21], presented the Hexad user type of
model. The Hexad is a gamification user types model
created to capture the users' motivations and different
styles of interaction with gameful systems.
This model proposed the following six user types:
1) Philanthropists are motivated by intent.
Without seeking a reward, they are altruistic
and eager to share.
2) Socializers are inspired by togetherness. They
want to communicate and build social relations
with others.
3) Free Spirits: autonomy and liberty empower
them to express themselves and behave without
external interference. Inside a system, they like
to produce and explore.
4) Achievers: competence motivates them. By
completing tasks, they strive to advance within
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
5
Volume 18, 2023
a system or show themselves by solving hard
challenges.
5) Players: external rewards or bonuses inspire
them.
6) Disruptors: the creation of transformation
influences them. They prefer to test the limits
of the system and disturb the system to force
negative or positive effects, either directly or
throw others.
5 Related Work
In universities and schools, the student data size is
increasing every year. So, we need to analyze this
data and extract useful information using adaptive
learning and educational data mining techniques.
In this section, we will list many research papers
that used data mining in education to solve the e-
learning challenges and how they solved these
problems.
One of the challenges faced in the educational
classroom is that every teacher needs to make the
student's assessment as a paper on the university, but
Orlando, etl , [24], solve this problem by building a
new system for evaluating students in undergraduate
courses to let students interact with him through this
system.
The authors, [24], designed an evaluation model
named ‘Leonardo’ to evaluate the student's work over
the system and this system didn’t remove the teacher
from evaluating the student and acquiring the
knowledge over time.
The system lets the student take many quizzes and
save the grades with some features The system
controls the assessment like timed questions and how
many questions are selected for the students from the
chapters they have learned on the system. Based on
the student’s questions and grades, he suggests a new
material focus on the question chapter he gets low
grades.
Figure 4 shows how this system works and
interacts with the student. One of the defects of this
system is that he didn’t save the student's behavior
after logging out from the system. Every time a
student logs into the system, he will be assigned the
same material as a cold start and a new user to the
system.
Fig. 4: Evaluation and profiling modules interact.
Lennart E. Nacke. Etl, [19], built a model called
BrainHex and used surveys and questionnaires to
classify users based on the results. This survey was
conducted for more than 50,000 players. This system
is like a game and classifies the user's type up to
seven learner styles: Daredevil, Mastermind,
Survivor, Seeker, Conqueror, Socialiser, and
Achiever.
They collect some demographic information
about the players and users to build relationships
between players personality types and BrainHex
model archetypes. A lot of researchers used this
paper to build a gamification e-learning model for
students to increase engagement with online
education.
Evgenia Baranova, etl, [25], one of the important
parts of the Russian Federation is to digitalize the
economy, and education is an important part of this
strategic goal.
The researcher used a digital education
environment to analyze the data generated from the
system to calculate the correlation between the online
structure data, educational programs, students'
behavior, and performance in the system.
They collect their data over 10 years from the
university on various aspects of the educational
process. calculations were made to find the
conclusions about the nature of relationships between
the selected features (the number of activities on the
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
6
Volume 18, 2023
course, the participation of students in the course
chat(s), the final assessment results for the course,
the time student spent on the course) obtained as part
of the student's work in the Moodle distance learning
system (Moodle DLS) and the results of the midterm
assessment.
Authors find that the activities may help in
measuring various disciplines and preparing
recommendations for updating educational resources
to improve their quality. Just like when the student
spends less time online and has a low grade on the
midterm and his activities on the system are low his
final grade will be low, and the system measure all
these factors and find which students are likely to fail
in this course to send a notification to the teacher to
follow these students before the end of the semester.
Chrysalis, etl, [6], propose innovative adaptive
learning based on two sources: the student and user
personalization information and their learning behavior
or style.
Firstly, to determine the first user behavior the
system suggests a test to the user that asks about his
learning style, based on these test results the system
will store this result and adjust or adapt the learning
material according to his style.
Secondly, they also used The Machine Learning
Management (MLM) component/module to access
the dataset that is formed from the Query Results
Management component (QRM) and for feeding it to
the actual machine learning algorithm.
This use case discusses how the system works. A
user (from here on known as learner 1) logs into the
system for the first time. Following that, learner 1 is
given an introductory Test to determine his or her
level of understanding. Based on this, the algorithm
may suggest some further study links, tutors to
contact, or even additional examinations. Let's
pretend that learner 1 chooses Topic 1 and takes Test
T1 after this initial phase. He gets a perfect score of
30. Based on the score obtained (and assuming a
performance criterion of 40 or 50), it is clear that
learner 1 requires additional support and assistance in
Topic 1. The experiments were performed with 300
learners to show the impact of learning styles on
learners’ preferences in this system.
Élise Lavoué, etl , [26], built an automatically
adapting gaming system for learning environments
This system uses gamification ideas with the learning
system to help increase users' motivation while using
this system. They get help from experts to build this
gamification system to build their website and the
user's adaptation model for the French language's
spelling and grammar learners The project, named
Project Voltaire, was developed by the Woonoz
company, which specializes in memorization
software.
They announce their project and ask anyone to
make a survey and fill in their data if they like to take
and try these experiments via their websites. 266
participants were asked to try this website and follow
these experiments to the end. In the second step, they
divide the participants into three groups randomly:
the first group will engage with the fully adaptive
gamification system and the second group has fewer
features of an adaptive system and the third group
has no gamification or adaptation on their system or
learning material.
After dividing the users into groups (the first and
the second group), they will build the user's profile
system initialized through the BrainHex survey that
remains identical during the learning activity time.
For three weeks, they first cluster users using this
survey and develop five gaming features
corresponding to different player types in the system.
Then the adaptive system will display the learning
material based on the experts' gamified learning
material they suggest from the beginning.
They make analyses and study the experimental
results. They use the user's time spent on the system
and the learning environment. They find the results
for the survey of enjoyment and gamification
environments has the greatest number of engaged
users than another non-gamified system. They should
make a questionnaire or a survey during the learning
process to know if the users enjoy this learning style
or not, if they need to complete these experiments at
the end of the three weeks, they find several learners
didn’t complete the survey and the learning
materials. Also, they talk about how an incremental
construction of the user profile could alleviate the
issue of the preliminary BrainHex questionnaire.
Ramlah Mailok,etl, [27], in their need to develop
digital children's games players aged 8–10 years by
using the brainHex model survey in their study
involving 214 Malaysian children. They need to
answer a question if there’s a relation between the
brainhex survey results and differences based on age
or gender. After their experiments, they found that
achiever, daredevil, and conqueror emerged as the
most dominant characteristics of all children at these
ages, and all developers of digital games can help
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
7
Volume 18, 2023
players, notably young players, or children, to
develop sound thinking skills.
Jalal Nouri, etl, [28], The flipped classroom is
used in this research topic as one of the learning
methods because it’s one of the pedagogical methods
that finds the highest combinations of physical
lectures and digital learning environments, like
blended learning environments. The Flipped
classroom is based on recording a video for all
lectures and uploading all learning materials digitally
on a website and finding the interaction between the
students and these materials. The interaction between
the students and the teacher is achieved by using
discussion forums and digital quizzes.
Jalal Nouri, et al, [28], applied the system using
255 students. Students at Stockholm University who
were given the course on research methodology in
autumn 2017 needed an analysis step, to do more
tasks as a group, and make a physical lecture to
measure the understanding of this uploaded material
with the teachers. This system also uses more
prediction methods to find how many students may
have passed or failed the course and detect the
students’ performance on this course using many
techniques and machine learning methods like (i.e.
Neural Networks, Naive Bayes, Random Forest,
kNN, and Logistic regression). Additionally, the
system takes into consideration the number of clicks
on the material for every student and predicts based
on his clicks on the system if he will pass or fail at
the end of the course.
Author's experiments finds that the KNN
prediction method has the highest accuracy for the
results if passed or failed students at the end of the
course, and if the students have less than 1355 clicks
on the system, they fail in the course because they
didn’t study the material well enough and their
results on the digital quizzes are less than 57 percent
of the total marks. Finally, the flipped classroom was
a good method for training students on the material
before making a discussion and applying the training
task with groups in the physical lectures, and it had a
great impact on students in the learning methods.
Siti Nurul Mahfuzah Mohamad,etl, [29], This
research idea builds a learning environment for
students based on their intelligence. Gamification is
used in this learning environment and makes the
ending and studying of the learning material like a
game. If you finish all tasks and assignments early,
students will get more marks and grades. All the
time, students must keep straggling and survive to
gain more marks to get the rewards and not die and
fail at the end of the course. This gamification
learning method will develop students' critical
thinking skills, increase their ability to work in a
group, and allow students to give marks and grades
for some assignments in the course, adding to the
final marks.
Experiments in this learning method were applied
to two groups of undergraduates from two TVET
institutions in Malaysia; group 1 has 36 students and
group 2 has 34 students studying the multimedia
course. There are five phases involved in this study:
1) Determining Student Intelligence by taking a
quiz and selecting which level to start the
course.
2) Defining Learning Goal: design the learning
material more interactive and make
assessment creative and design activities with
proper game elements.
3) Structuring Learner Experience: they must
select the more appropriate game types that
can be used with this subject and helpful to let
students enjoy learning time all this semester.
4) Analyze Suitable Game Elements: the system
interface and usability for all students selects
12 game elements to be added to the system
list in Table 2.
5) Heutagogy Design Process: train students on
how to use the technology and how to solve
the assignments because this course is a
multimedia system that teaches students how
to make an online presentation, poster design,
creative platform, problem-based learning,
how to connect with Adobe Education
Exchange, drag and drop activities, and how
to solve crossword puzzles.
Using The gamified E-Learning site (OMIG),
[29], is also embedded into LEARN, as well as
linked to OpenCourseware (OCW) and Eg-MOOC.
The new idea in OMIG, students can go to the next
level in the learning process based on student
intelligence and the results from the first MI
(multiple intelligence test). Authors publish this link
to be public to allow students to learn online
http://onmitt.net/omig/index.html,https://www.flipsn
ack.com/995DFECF8D6/a-online-delivery.html.
At the end of this multimedia gamified course,
they find that successful gamification needs to adapt
to students’ intelligence and provide suitable
teaching materials to master their skills. Creative
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
8
Volume 18, 2023
educators can make intelligent materials more
attractive to students with a simple way to learn.
Table 2. Top Twelve Game Elements
No
Interaction
Dimension
Gamification
Elements
1
Virtual Goods
Wally’s Games
Memory Game
Check Points
2
Rewards
3
Redeemable Points
Skill Points
Peer grading
Peer Emoticon
Feedback
Team Leaderboard
Albert a.shawky,etl, [30], [31] ,built a technique
to enhance the effectiveness of adaptive e-learning
solutions and to create an effective online
recommendation system, to correlate the student
variables such as knowledge level and learning style.
This can be resolved by creating a new learner-
centered model, [30], that incorporates learner style
and knowledge level. Two efficient adaptive e-
learning models are proposed. The first model for
recommending materials is based on the prediction of
the learner style using a questionnaire and adapted
fuzzy c mean (FCM) , and the second one is for
recommending materials based on a prediction of the
learner style using a questionnaire and Learning
Management Platform (LMP) score. In this model,
questions and learning materials are mapped to each
LMP. This model is a multi-level model that
calculates the LMP values as well as stores the
marks.
This system results show that the Adaptive fuzzy
c mean system enhances the performance of the
learner knowledge level by using the questionnaire
and the adaptive FCM model has achieved the best
performance at 88.7%. Moreover, the second
contribution called an adaptive e-learning
recommender model based on Knowledge Level and
the Learning Style (AERM-KLLS) achieves the
highest accuracy at 90.97.
Przybylski et al, [32], This researcher finds out
how video games meet these psychological demands,
by proposing a motivating model based on Self-
Determination Theory (SDT).
Consider the following scenario: Obtaining
feedback and demonstrating improvement will
reward you. Then providing options for methods and
possibilities will satisfy the sense of competence.
Autonomy and competitiveness in the leaderboard
and forum cooperation will meet the needs of the
participants.
Reem S. Al-Towirgi, etl, [33], Students used a
gamification system to learn the data structure course
to evaluate students while using the physical teaching
methods in the classroom and to evaluate the
students' grades while using this gamification data
structure course. They built a gamified course using
the Moodle platform and used a more attractive user
interface, themes, more fonts, and colors to make the
system more enjoyable to the students. They used
many levels in this system like Avatar, Options in
selecting topics, Options in determining badges,
Progress, Challenges, feedback, and Leaderboards.
To satisfy the competence needed according to Self-
Determination Theory (SDT), satisfying these three
psychological needs will enhance the self-motivation
of the students.
They conducted exterminates for 40 students of
the data structure course. Firstly, they made an
introduction to the course and illustrated how to use
the system to get more points.
Firstly, students log in to a gamified course and
take a pre-test to know the student’s knowledge of
the course and cover some course topics. Students
will navigate the course and study the material via a
gamification system. At the end of the course,
students must take an online test to measure the
effectiveness of the gamification on their learning
outcome. Also, students were asked to fill out an
online questionnaire about their feedback regarding
their experience. A questionnaire will be used to
measure student engagement. The final students'
results and satisfaction with the overall gamification
learning course have got high results compared to the
teaching in the classroom.
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
9
Volume 18, 2023
6 Adaptive Learning with Facial
Recognition
Uğur Ayvaz, etl, [34], E-learning has the advantage
of providing more flexibility while we can discuss
material online at any time with our students and
solve the problem of student capacity. One of the
main problems in e-learning is that we didn’t use it
face-to-face and didn’t see the student reactions
while we discussed the material and topics. They
build a system used to take an image of the students’
emotions while they use Skype online meetings. The
teacher discusses how the material system will save
these screenshots based on the students’ emotions.
They turn this image to give feedback to the teacher
instantaneously. If students understand this material
or not, whether they are happy, angry, or neutral,
many different emotion types are used in their
systems.
Uğur Ayvaz, etl, [34], build a system named a
FERS to send a report about the students’ emotions
to the teacher this system used SVM (support vector
machine) to make a classification based on students
matching emotions and this system has more accurate
prediction has 98% percentage to the real emotion
saved before starting these experiments.
Duong Thang Long, [35], in this research paper,
the authors need to solve the problem of how to
identify the students in a learning management
system. He attends the course and logs into the
material online as a part of e-learning systems. The
authors propose a new model based on convolutional
neural networks (CNN) for human face recognition
problems that has 5 convolutional neural layers
(CONV) and 2 fully connected neuron layers (FC).
This model can detect the student's images of
different complexity and different light, like dark
images, and it can crop the noise around the student's
face to make more accurate results to identify the
user. They train this model on different datasets like
the AT&T dataset (also called ORL), which was
created by the AT&T Laboratory at Cambridge
University, in 2002. It includes 400 images of 40
people, with 10 different attitudes for each person.
The Yale dataset was created by the Computer
Control and Visual Center at Yale University. It
consists of 165 images taken from the front and on
the multi-level of 15 different people.
The LFW dataset has a diverse number of images,
ranging from 5 to 530. We just use people who have
20 or more face images. It is the so-called LFW20.
So, LFW20 has 3,023 images of 62 people.
Their dataset, [35], was collected in an online
class with 24 students. It has 1,005 face images, from
the lowest number of 5 images to the largest number
of 222.
This model is integrated with an LMS (learning
management system). After the student logs in to the
system with his user id and password, the system
takes an image of the student using the CNN model
and detects whether this account belongs to him or
not by following these steps:
1) Open the client’s camera to capture images of
students. This activity is integrated with LMS on the
client side for every student.
2) Preprocess captured images to get face images
from the client’s camera.
3) Recognize face images, [36], to get the ID of
students or ‘unknown’ send notifications to LMS for
announcing and monitoring the whole learning time
of students.
Mohammed Megahed, etl, [37], The importance
of adaptive learning environments to track the
student’s response while studying the material if they
can go to the next level or need more practice at this
level again. more systems of adaptive learning
methods failed to capture the emotions of the
students learning while studying the course. They
build a new system to take screenshots of the
students while studying the materials and taking
quizzes by using an integrated system with CNN
(conventional neural network) and fuzzy C-Means to
cluster the students with the best knowledge level
and their ability to study the next chapter.
In the proposed system Mohammed Megahed, etl,
[37] , takes the responsibility for the exam and timed
test questions and gives more different emotions to
match his emotions and his grades in the system.
Also, taking a screenshot of the students' learning
time on the system can give the educator or instructor
a brief about the student’s grades expected at the end
of the course. The system's main components are
facial expression recognition, test, and exam
manager, learning modeler, and fuzzy inference
system. The importance of a fuzzy inference system
is that it receives this data from the learner model to
cluster students to the next learning levels or give
more time to study the same topic again with more
materials. The experimental results based on the
corpora of 12 learners contain 72 learning activities
and 1735 data points of distinct emotional states.
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
10
Volume 18, 2023
7 New Proposed Model
● (ILOS ADAPTIVE LEARNING WITH THE
EMOTION-BASED SYSTEM)
We need to build an adaptive learning system to
solve new problems and challenges today with the
higher education about how to discuss the material
with the students according to intended learning
outcomes (ILOs), build a material to find and
determine the knowledge level of students, and how
the teacher will demonstrate this course to let
students understand materials and which practical
skills they will learn from the course.
This new proposed model will be adapted to the
students' learning styles by making course materials
fully engaged with pdfs, videos, and PowerPoints
lecture materials. They will need to find their path by
solving assignments and quizzes and how to get more
points by solving assignments early and on the
requested time. There are many models used to
determine the student’s learning style based on his
activity on the system, like:
1. Kolb Learning Style Model
2. VARK Learning Style Model
3. Gregorc Learning Model
4. Hermann Brain Dominance
5. MAT Learning Model
6. Felder-Silverman Learning Style Model
7. Honey Mumford Model
We will use VARK. It will be suitable for our
learning materials VARK (It stands for visual,
auditory, reading/writing, and kinesthetic learning
styles). This model states that every learner
experiences learning through any one of these
processes.
Our system will classify students based on their
learner styles and suggest suitable learning materials
like videos or reading textbooks or taking more
practice with sheets.
Our material divides according to the ILO's the
student will learn from every lesson he will study.
The main target of the system is how to make an
exam that satisfies the ILOs of the course. Our
system will choose several questions at the final
assessment based on the question bank we build
based on the learning outcomes from every lesson
they study and which knowledge we need to measure
through different assessment methods. Auditory
learners frequently converse with themselves.
Reading and writing assignments may be a challenge
for them. They typically do better talk to a colleague
or listening to what was said on a tape recorder.
To adapt this learning technique to teaching
adaptive learning material, we follow these steps:
● We'll start with new material with a quick
overview of what's the outcome.
● Finish with a summary of what you've learned
thus far. "Tell them what they're going to
learn, teach them, then tell them what they've
learned," goes the ancient adage.
● Use the Socratic technique of lecturing by
interrogating students to obtain as much
information as possible from them, then
filling in the gaps with your knowledge.
● Incorporate auditory activities like
brainstorming, buzz groups, and so forth.
● Create an internal dialogue between you and
your students’ using chats and forums.
Visual learners have two sub-channels—linguistic
and spatial. Learners who are visual-linguistic prefer
to learn through written language, such as reading
and writing assignments. Even if they do not read it
more than once, they recall what has been written
down. Even if they do not read it more than once,
they recall what has been written down. They prefer
to take notes and will pay more attention to lectures
if they observe them. Visual-spatial learners typically
struggle with written language and perform better
with charts, demonstrations, movies, and other visual
materials. They can quickly imagine faces and places
using their imagination, and they rarely become
disoriented by new learning methods. To incorporate
this learning technique into the adaptive system, we
add this feature:
● Include graphs, charts, pictures, or other
visual aids in your presentation.
● Include reading and note-taking by writing
comments like outlines, idea maps, agendas,
and handouts.
● Leave blank spaces in handouts for taking
notes and making comments.
● Ask them questions to keep them aware and
allow space for discussion in the chat groups
on the system.
● Use flip charts to depict what will happen next
and what has already happened.
● Highlight crucial points to indicate when it's
time to take notes.
Kinesthetic learners do best while touching and
moving. You're at your finest. There are two sub-
channels: kinesthetic (movement) and tactile (touch).
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
11
Volume 18, 2023
If there is little or no external stimulus or movement,
they will lose concentration. They may wish to take
notes while listening to lectures to keep their minds
active. They prefer to scan the content first before
focusing on the specifics when reading (getting the
big picture first). Color highlighters are commonly
used, and they take notes by drawing images, and
diagrams. To incorporate this learning technique into
the adaptive learning system, we follow these steps:
● Engage learners in activities that get them up
and moving.
● Play music during activities when it is
acceptable.
● Highlight essential points on flip charts or
whiteboards with colored markers.
● Allow for frequent breaks for online
sessions (brain breaks).
● Make sure you have highlighters, colored
words, and sentences on your presentations
and graphics on videos recorded.
● Lead students through visualization of
difficult problems.
Our new system records the emotions, [38], [39],
of the students while taking a quiz to make a report to
the instructor on which part these students need more
practice and to find the accuracy of the system
between the system suggestion and the results of
every question passed on this quiz.
Finally, our new system will solve more adaptive
learning methods challenges and will let students
engage with the system easily.
8 Experimental Results
Today, all schools and universities build an online
learning system to help students learn and study all
course materials remotely. It also helps teachers mark
the students’ assignments, quizzes, and exams online.
To test our proposed model, we need to build the
course materials to match with the course ILOs, and
the materials must have videos, pdf files, and ppt to
detect the learner style model. So, we test our
proposed system on real course students at a
university.
The system was online for students for three
months from the start of the term to the end of it.
This system is built using the NEO website.
The system has 310 learners studying the online
course, Operations Research. The first week: starts
by sending the students a link to our website and
teaches them how to join this course via invitation
through emails; collects data about the learner to
initialize and create the learner profile. It also creates
the domain model to help us discover the learner
styles of students after tracking their activity on this
website.
Table 3. Teaching and Learning Methods
Lecture
Brainstorming
Discussions
Tutorials
Problem solving
Laboratory &
Experiments
Research and Reports
Role playing
Workshops
Projects
Modeling and
Simulation
To build our new proposed models, we use the
course ILOS first, based on the course specification
and road map for the knowledge needed to be learned
by students during this semester. The Teaching and
Learning Methods as in table 3 and the Assessment
Methods as in table 4 suitable for every ILOs
(Knowledge & Understanding, Intellectual skills,
Professional skills and General Tran. Skills) from the
course.
Table 4. Assessment Methods
Written Exam
Practical / Exercise
Exam
Quizzes
Term Papers
Assignments
After students register on the website and add
their demographic data, we track the student’s
activity on our website to detect the student’s
learning style based on the VARK Model.
To Cluster students by K- means centroid based
on their activity we used word net firstly to make
nouns and sentences that define all these nouns and
keywords match to all learner styles and keywords.
Building a semantic map to match between these
keywords to cluster students into 4 groups Visual,
Aural, Read and write and kinesthetic figure 5 has all
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
12
Volume 18, 2023
activity for students to match with any learner style
like visual students like assignments, lesson,
workshops, forums and wiki.
Fig. 5: VARK Learning style
We will list some of our system screenshots
showing the model website contains materials and
quizzes and assignments. Figure 6 shows the home
page of the course. It contains 10 lectures, all
different materials, has videos and pdf files
assignment to let students answer these questions and
try to solve problems.
Figure 7 shows The content for lecture 2 with
video and pdf file for this lecture, Figure 8 shows the
assignment content, start time , end time , grade ,
number of students submitting this assignment within
time and number of students delayed in submitting
this task. Figure 9 shows the content for the quiz
questions created based on the ILos of the course.
Fig. 6: Home page
Fig. 7: Lecture 2 content
Fig. 8: The assignment content
Fig. 9: The content of the quiz question.
Our model divides our students based on their
activity. We found that the majority of students were
visual learners (45%) and the rest of the sample
students were kinesthetic and auditory, that is 30%
and 25% respectively. Figure 10 represents the
distribution of sample students’ learning styles.
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
13
Volume 18, 2023
Fig. 10: learner style distributions
Finally, after the semester ends, we compare the
students’ grades while studying in our system to the
previous year. Students study all lectures for this
course physically on campus. We found that the
grade scores for students increased this year and
there was high satisfaction from the students' survey
questionnaire.
Figure 11 shows the grades for every learner
style. In this model we found that visual students
have the highest grades followed by kinesthetic and
the lowest grades auditory.
Fig. 11: Learner style and final grades
9 conclusion
Adaptive learning systems have the most important
topics today. All researchers do a lot of systems to
increase student engagement by using gamification to
make lessons like a game with levels that must pass
the first level and collect more badges or see the
leaderboard list then can go to the next level. Other
systems are used to classify students based on their
activity on the site to detect their learner style and
suggest the material. According to this report, many
data mining techniques are used like clustering k-
means, and classification like fuzzy c-means to
predict results more accurately. Another system
needs to detect the users’ emotions while studying
and while taking quizzes to predict their grade and
which topic they need to study again.
Our new proposed system of ILOs adaptive
learning with an emotion-based system will detect
the student’s learning style according to the student’s
activity on the site and then cluster students based on
this style. The new idea is how to match this material
with the ILOs of the course (Intended learning
outcomes) as one of the most requested in learning
today. The final step of the system is making the quiz
match more questions based on the ILOs of the
course engaged with monitoring the emotions while
the students take the quiz to predict their grades and
topics, he needs to study again according to the
lowest grades.
According to this study, we found that visual
learners like essays and demonstrating a process
assessment; auditory learners like writing comments
on lectures and like oral exams; and kinesthetic
learners like MCQ questions and complete questions.
So, we need to update our material to match all the
learner-style assessments to let all students get high
grades with the learning outcomes and preferred
assessment methods.
future work, what about adding facial emotion
tracking while studying this course material in our
system and measuring the accuracy of prediction
from this system and measuring the student's
performance and the difference between studying the
course in the classroom or using this adaptive
learning system.
References:
[1] S. Bal Gupta, R. Kumar Yadav, and A.
Professor, “Analysis of Popular Techniques
Used in Educational Data Mining,” Jul. 2007.
45%
25%
30%
students Learner style
Visual
Auditory
Kinesthetic
75
50
70
0
10
20
30
40
50
60
70
80
Visual Auditory Kinesthetic
Final Grades
Visual Auditory Kinesthetic
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
14
Volume 18, 2023
[Online]. Available:
https://www.researchgate.net/publication/3444
39762
[2] L. Gabriel et al., “Data mining techniques
applied in educational environments: Literature
Review,” 2018. DOI: Digital Education
Review - Number 33, June 2018.
[3] M. Hussien Mohamed, M. Khafagy, M. Hasan
Ibrahim, and M. Helmy Khafagy, “From Two-
Way to Multi-Way: A Comparative Study for
Map-Reduce Join Algorithms LocalitySim
View project Fatma Omara’s Lab View project
From Two-Way to Multi-Way: A Comparative
Study for Map-Reduce Join Algorithms,” 2018.
[Online]. Available:
https://www.researchgate.net/publication/3259
57573
[4] M. H. Mohamed and M. H. Khafagy, “Hash
semi cascade join for joining multi-way map
reduce,” in IntelliSys 2015 - Proceedings of
2015 SAI Intelligent Systems Conference, Dec.
2015, pp. 355–361. DOI:
10.1109/IntelliSys.2015.7361167.
[5] L. F. Daghestani, L. F. Ibrahim, R. S. Al-
Towirgi, and H. A. Salman, “Adapting
gamified learning systems using educational
data mining techniques,” Computer
Applications in Engineering Education, vol.
28, no. 3, pp. 568–589, May 2020, DOI:
10.1002/cae.22227.
[6] C. Chrysoulas and M. Fasli, “Building an
adaptive E-learning system,” in CSEDU 2017 -
Proceedings of the 9th International
Conference on Computer Supported Education,
2017, vol. 2, pp. 375–382. DOI:
10.5220/0006326103750382.
[7] R. S. Al-Towirgi, L. F. Daghestani, and L. F.
Ibrahim, “Data Mining and Gamification
Techniques in Adaptive E-Learning: Promises
and Challenges,” 2018.
[8] M. H. Mohamed, M. H. Khafagy, and M. H.
Ibrahim, “Hash Semi Join Map Reduce to Join
Billion Records in a Reasonable Time,” Indian
Journal of Science and Technology, vol. 11,
no. 18, pp. 1–9, May 2018, doi:
10.17485/ijst/2018/v11i18/119112.
[9] X. X. Suhirman, J. M. Zain, and T. Herawan,
“Data mining for education decision support: A
review,” International Journal of Emerging
Technologies in Learning, vol. 9, no. 6, pp. 4–
19, 2014, DOI: 10.3991/ijet.v9i6.3950.
[10] M. Alsuwaiket, A. H. Blasi, and K.
Altarawneh, “Refining Student Marks based on
Enrolled Modules’ Assessment Methods using
Data Mining Techniques,” Engineering,
Technology & Applied Science Research, vol.
10, no. 1, pp. 5205–5210, Feb. 2020, DOI:
10.48084/etasr.3284.
[11] J. L. Rastrollo-Guerrero, J. A. Gómez-Pulido,
and A. Durán-Domínguez, “Analyzing and
predicting students’ performance by means of
machine learning: A review,” Applied Sciences
(Switzerland), vol. 10, no. 3. MDPI AG, Feb.
01, 2020. DOI: 10.3390/app10031042.
[12] T. Y. Tran and B. Lam To, “Educational Data
Mining for Supporting Students’ Courses
Selection,” 2019.
[13] A. U. Khasanah and H. Harwati, “Educational
data mining techniques approach to predict
student’s performance,” International Journal
of Information and Education Technology, vol.
9, no. 2, pp. 115–118, Feb. 2019, DOI:
10.18178/ijiet.2019.9.2.1184.
[14] M. Choe, J. Yoo, G. Lee, W. Baek, U. Kang,
and K. Shin, “MiDaS: Representative
Sampling from Real-world Hypergraphs,” Feb.
2022, DOI: 10.1145/1122445.1122456.
[15] N. Aldhafferi, A. Alqahtani, and K. Sultan,
“Educational data mining for enhanced
teaching and learning,” Article in Journal of
Theoretical and Applied Information
Technology, vol. 31, no. 14, 2018, [Online].
Available:
https://www.researchgate.net/publication/3268
49608
[16] T. Toivonen, I. Jormanainen, and M.
Tukiainen, “Augmented intelligence in
educational data mining,” Smart Learning
Environments, vol. 6, no. 1, Dec. 2019, DOI:
10.1186/s40561-019-0086-1.
[17] D. Topalli and N. E. Cagiltay, “Improving
programming skills in engineering education
through problem-based game projects with
Scratch,” Computers and Education, vol. 120,
pp. 64–74, May 2018, DOI:
10.1016/j.compedu.2018.01.011.
[18] Yang, Tzu Chi & Hwang, Gwo-Jen & Yang,
Stephen. Development of an Adaptive
Learning System with Multiple Perspectives
based on Students' Learning Styles and
Cognitive Styles. in Journal of Educational
Technology and Society. Vol.16.pp. 185-200
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
15
Volume 18, 2023
[19] L. E. Nacke, C. Bateman, and R. L. Mandryk,
“BrainHex: A neurobiological gamer typology
survey,” Entertainment Computing, vol. 5, no.
1, pp. 55–62, 2014, DOI:
10.1016/j.entcom.2013.06.002.
[20] R. Daud, S. Salam, S. Nurul, M. Mohamad,
and A. M. Yusoff, Gamification Leaderboard
model based on player type to increase student
engagement. 2018. [Online]. Available:
https://www.researchgate.net/publication/3389
90921
[21] J. T. Avella, M. Kebritchi, S. G. Nunn, and T.
Kanai, “Learning Analytics Methods, Benefits,
and Challenges in Higher Education: A
Systematic Literature Review.”
[22] S. Bennani, A. Maalel, and H. ben Ghezala,
“Adaptive gamification in E-learning: A
literature review and future challenges,”
Computer Applications in Engineering
Education. John Wiley and Sons Inc, 2021.
DOI: 10.1002/cae.22477.
[23] É. Lavoué, B. Monterrat, M. Desmarais, and S.
George, “Adaptive Gamification for Learning
Environments.”
[24] O. Belo, J. Coelho, and L. Fernandes, “AN
EVOLUTIONARY SOFTWARE TOOL FOR
EVALUATING STUDENTS ON
UNDERGRADUATE COURSES,” 2019.
[25] E. Baranova, G. Shvetsov, and T. Noskova,
“Educational Data Mining Methods for the
Analysis of Student’s Digital Footprint *.”
[26] É. Lavoué, B. Monterrat, M. Desmarais, and S.
George, “Adaptive Gamification for Learning
Environments,” IEEE Transactions on
Learning Technologies, vol. 12, no. 1, pp. 16–
28, Jan. 2019, DOI:
10.1109/TLT.2018.2823710.
[27] Ramlah Mailok, Mashitoh Hashim, Haslina
Hassan, and Noriza Nayan, “The Personality of
Digital Games Players among Children based
on the Brain Hex Model 402,” International
Journal of Innovative Technology and
Exploring Engineering (IJITEE) , vol. 8, no.
Issue-7S2, 2019, DOI: 10.3390/j1010008.
[28] J. Nouri, M. Saqr, and U. Fors, “Predicting
performance of students in a flipped classroom
using machine learning: towards automated
data-driven formative feedback,” 2019. DOI:
SYSTEMICS, CYBERNETICS AND
INFORMATICS VOLUME 17 - NUMBER 2 -
YEAR 2019.
[29] S. N. Mahfuzah Mohamad, M. A. Mohd
Salleh, M. H. Abdul Hamid, Bl. K. Mei Sui,
and C. K. N. Che Ku Mohd, “Adaptive
Learning Strategies with Gamification to
Enhance Learning Engagement,” Indian
Journal of Science and Technology, vol. 12,
no. 31, pp. 1–8, Aug. 2019, DOI:
10.17485/ijst/2019/v12i31/146871.
[30] A. S. Aziz, R. A. El-Khoribi, and S. A. Taie,
“AFCM MODEL TO PREDICT THE
LEARNER STYLE BASED ON
QUESTIONNAIRE AND FUZZY C MEAN
ALGORITHM,” Journal of Theoretical and
Applied Information Technology, vol. 31, no.
2, 2021, [Online]. Available: www.jatit.org
[31] A. S. Aziz, R. A. El-Khoribi, and S. A. Taie,
“ADAPTIVE E-LEARNING
RECOMMENDATION MODEL BASED ON
THE KNOWLEDGE LEVEL AND
LEARNING STYLE,” Journal of Theoretical
and Applied Information Technology, vol. 99,
p. 22, 2021, [Online]. Available: www.jatit.org
[32] A. K. Przybylski, C. S. Rigby, and R. M. Ryan,
“A Motivational Model of Video Game
Engagement,” Review of General Psychology,
vol. 14, no. 2, pp. 154–166, Jun. 2010, DOI:
10.1037/a0019440.
[33] L. F. Daghestani and L. F. Ibrahim, “Increasing
Students Engagement in Data Structure Course
Using Gamification,” International Journal of
e-Education, e-Business, e-Management and e-
Learning, vol. 8, no. 4, pp. 193–211, 2018,
DOI: 10.17706/ijeeee.2018.8.4.193-211.
[34] U. Ayvaz, H. Gürüler, and M. O. Devrim,
“USE OF FACIAL EMOTION
RECOGNITION IN E-LEARNING
SYSTEMS,” Information Technologies and
Learning Tools, vol. 60, no. 4, p. 95, Sep.
2017, DOI: 10.33407/itlt.v60i4.1743.
[35] D. T. Long, “A lightweight face recognition
model using convolutional neural network for
monitoring students in e-learning,”
International Journal of Modern Education and
Computer Science, vol. 12, no. 6, pp. 16–28,
2020, DOI: 10.5815/ijmecs.2020.06.02.
[36] M. Hussien Mohamed, M. Helmy Khafagy, H.
Elbeh, and A. Mohamed Abdalla, “Sparsity
and Cold Start Recommendation System
Challenges Solved by Hybrid Feedback,” 2019.
[Online]. Available: http://www.irphouse.com
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
16
Volume 18, 2023
[37] M. Megahed and A. Mohammed, “Modeling
adaptive E-Learning environment using facial
expressions and fuzzy logic,” Expert Systems
with Applications, vol. 157, Nov. 2020, DOI:
10.1016/j.eswa.2020.113460.
[38] M. Hussien Mohamed, M. Hasan Ibrahim, M.
Khafagy, and M. Helmy Khafagy, “Two
recommendation system algorithms used SVD
and association rule on implicit and explicit
data sets SEED4NA-SDI and EO Education
and Training for North Africa-Erasmus+ View
project BeesWax View project Two
Recommendation System Algorithms Used
SVD And Association Rule On Implicit And
Explicit Data Sets,” INTERNATIONAL
JOURNAL OF SCIENTIFIC &
TECHNOLOGY RESEARCH, vol. 9, p. 1,
2020, [Online]. Available: www.ijstr.org
[39] M. Hussien Mohamed, M. Helmy Khafagy, M.
Hasan Ibrahim, K. Elmenshawy, and H. Rizk
Fadlallah, “MUSIC RECOMMENDATION
SYSTEM USED EMOTIONS TO TRACK
AND CHANGE NEGATIVE USERS’
MOOD,” Journal of Theoretical and Applied
Information Technology, vol. 15, p. 17, 2021,
[Online]. Available: www.jatit.org
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
The authors equally contributed in the present
research, at all stages from the formulation of the
problem to the final findings and solution.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
No funding was received for conducting this study.
Conflict of Interest
The authors have no conflicts of interest to declare
that are relevant to the content of this article.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2023.18.1
Marwa Hussien Mohamed, Lamiaa F. Ibrahim,
Khaled Elmenshawy, Haitham Rizk Fadlallah
E-ISSN: 2224-2856
17
Volume 18, 2023