Content uploaded by Shatha Ghareeb
Author content
All content in this area was uploaded by Shatha Ghareeb on Jun 09, 2021
Content may be subject to copyright.
Abstract—In Abu Dhabi, there are many different education
curriculums where sector of private schools and quality assurance is
supervising many private schools in Abu Dhabi for many
nationalities. As there are many different education curriculums in
Abu Dhabi to meet expats’ needs, there are different requirements for
registration and success. In addition, there are different age groups
for starting education in each curriculum. In fact, each curriculum has
a different number of years, assessment techniques, reassessment
rules, and exam boards. Currently, students that transfer curriculums
are not being placed in the right year group due to different start and
end dates of each academic year and their date of birth for each year
group is different for each curriculum and as a result, we find
students that are either younger or older for that year group which
therefore creates gaps in their learning and performance. In addition,
there is not a way of storing student data throughout their academic
journey so that schools can track the student learning process. In this
paper, we propose to develop a computational framework applicable
in multicultural countries such as UAE in which multi-education
systems are implemented. The ultimate goal is to use cloud and fog
computing technology integrated with Artificial Intelligence
techniques of Machine Learning to aid in a smooth transition when
assigning students to their year groups, and provide leveling and
differentiation information of students who relocate from a particular
education curriculum to another, whilst also having the ability to
store and access student data from anywhere throughout their
academic journey.
Keywords—Admissions, algorithms, cloud computing,
differentiation, fog computing, leveling, machine learning
I. INTRODUCTION
N multicultural countries where there are many different
nationalities, there are different education curriculums
available to satisfy parents and students’ needs. Switching
from one curriculum to another has become a critical issue as
differences among curricula could create gaps in education
levels for students. Hence, it is difficult to assign the right
level initially for students when moving to a new curriculum.
The question is how we can make this transition more precise
and smoother with very minimum effect on students’
performance. There are many education curricula in Abu
Dhabi, and different education curricula have students starting
at a different age; for example, some education curricula (e.g.,
Philippine) start from 2.5 years old, whilst some others start
from 4 years old (e.g., India and Pakistan). According to
analyses of education curricula outcomes in Abu Dhabi and
Shatha Ghareeb is with the Liverpool John Moores University, United
Kingdom (e-mail: S.R.Ghareeb@2019.ljmu.ac.uk).
based on students’ results in high school level, most students,
who have transferred from one education curriculum to
another, have a gap in their education skills and knowledge
[1].
Although distributed computing paradigms (e.g. cloud
computing) and Machine Learning (ML) have been used a lot
“separately” in education for different purposes (e.g. sharing
contents and delivering lectures online in the case of cloud,
and predicting students’ performance using ML), they have
never been utilized jointly to perform leveling in multicultural
education countries. Therefore, there is an instant need for
having a united computerized system for automating the
leveling process for students when changing between the most
common education curricula. Nevertheless, there is a huge
ambiguity in this area, which usually creates conflicts between
parents, school, and Ministry of Education. This conflict has
obviously a negative impact on the students’ performance
when making the wrong decision.
This paper aims at proposing and developing a
computational framework applicable in multicultural
countries, such as UAE, in which multi-education curriculums
(i.e., UK and USA curricula) are implemented. The ultimate
goal is to aid in a smooth transition during admissions,
leveling and differentiation of students who relocate from a
particular education curriculum to another; and minimize the
impact of switching on the students’ educational performance.
II. CLOUD AND FOG COMPUTING IN EDUCATION
With the use of cloud, the world will become a classroom,
teaching process will be changed by applying e-learning.
Students can learn and teachers can teach form any place.
With use of cloud computing, students’ work can be set,
collected and graded online by teachers, while teachers can
also put the required resources for students online to use so
that students take responsibility of their own learning [2]. The
resources could include videos, documents, audio podcasts or
interactive images [3]. As long as there is internet connection,
students can access these resources via their personal
computer, smartphone or tablet. Sharing applications and
documents on the cloud, like Google apps provides better
opportunity for social lessons, this will enhance the
collaborative productivity between students, students can work
together if they are in the same room or in different places.
These collaborative tools are very helpful for teachers as well.
Many schools are following this currently, yet still there is a
long way to make sure that all schools are embracing it [4].
A Machine Learning Based Framework for Education
Levelling in Multicultural Countries: UAE as a Case
Study
Shatha Ghareeb, Rawaa Al-Jumeily, Thar Baker
I
World Academy of Science, Engineering and Technology
International Journal of Humanities and Social Sciences
Vol:14, No:12, 2020
1181International Scholarly and Scientific Research & Innovation 14(12) 2020 ISNI:0000000091950263
Open Science Index, Humanities and Social Sciences Vol:14, No:12, 2020 waset.org/Publication/10011632
In UAE a smart learning program started in 2012 named
Mohammad Bin Rashid smart learning program (MBRSLP).
This program will provide a new learning environment in the
national schools, with lunching of smart classes. Smart
education has become the new trend in educational field
globally. Intelligent technologies such as could computing,
Internet of things and learning analytics have developed the
emergence of smart education. These technologies focus on
how to capture learning data, analyse it and use it to improve
education [5].
A. Related Work
Chandra and Bora [6] did a research on the role of Cloud
Computing in education. The researcher came to a conclusion
that Cloud Computing will introduce a change in the way
teaching is provided to students by allowing teachers to focus
on teaching and research activities rather than on intricate IT
implementation. Chandra and Borah also stated that Learning
as a Service (LaaS) is likely to be a new way of cloud
computing education platform.
Khan et al. [7] presented a research on Education Cloud
Model and ways the cloud can be implemented in higher
education. The research shows that users (students, teachers,
parents, etc.) would be able to access different education cloud
services with use of IoT devices (laptops, smart phones, etc.).
Education Cloud Models can provide an extensive linked
structure between universities.
Rao [8] presented a new model for education as a service
(EaaS), which enables stakeholders (students, teachers,
parents, etc.) to overcome current challenges faced in
education in terms of communication and learning in general.
The suggested model demonstrates different services by the
cloud in the form of an Education Management System
(EMS).
III. DATA MINING AND ML IN EDUCATION
Data mining or Knowledge Discovery in Databases (KDD)
is the concept of finding the novel and useful information
from a huge amount of data [9]. Recently there is an increased
interest in using data mining in educational research,
Educational data mining (EDM) is focusing on developing
methods for finding the unique and important data which
come from educational setting and using them to enhance the
level of understanding about students and the setting [10].
There are many applications of EDM; one of the key areas
of applications which got an attention within the field is in
improving student models which provides information about
students’ level of knowledge, motivation, attitudes and other
characteristics [11]. The most essential theme in educational
software research is modeling students’ individual differences
for enabling software to react and deal with these differences
[12].
Most practical ML uses supervised learning. Supervised
learning is where we have an input variable known is (x) and
an output variable (y), and then we use an algorithm to learn
the mapping function from the input to the output [13].
Y=f(X)
The purpose of supervised learning is to make an
approximate of the mapping function close to accurate so that
when we introduce a new input data (x), we are able to predict
the output variable (Y) for that new data inserted. Supervised
learning is anticipated in find the patterns in the data which
can then be applied to an analytics process. [13]
There are many ways in which ML can enhance education
process in the coming future, which are:
1) Customisable learning experience
2) Student path prediction
3) Unbiased grading system
4) Overall feedback on both students’ and teachers’
performance
IV. IDENTIFIED GAPS IN CURRENT SYSTEMS
Fig. 1 Standard admission process followed by a school in Abu Dhabi
It has been seen that the education sector in general
underutilizes the advanced technologies and is not able to
proficiently increase its operational efficiency. Likewise most
of the students’ records are on papers inhibiting information
sharing. The admissions procedures as shown in Fig. 1 that are
World Academy of Science, Engineering and Technology
International Journal of Humanities and Social Sciences
Vol:14, No:12, 2020
1182International Scholarly and Scientific Research & Innovation 14(12) 2020 ISNI:0000000091950263
Open Science Index, Humanities and Social Sciences Vol:14, No:12, 2020 waset.org/Publication/10011632
currently used in schools across Abu Dhabi are not that
efficient in terms of time consumed and errors caused.
Previously, schools in Abu Dhabi have considered that the
year system is equal to the grade system, and therefore
students were assigned to the wrong year/grade groups.
Overall student admission, levelling, and differentiation is not
consistent throughout schools as some use their own levelling
criteria while some use external agencies.
V. P
ROPOSED
D
ISTRIBUTED
ML
B
ASED
F
RAMEWORK
The framework consists of three layers, intelligent decisions
layer (Stakeholders), fog layer and cloud layer, as shown in
Fig. 2, schools make intelligent decisions with the use of
different IoT devices (PC, Laptop, Smartphone, and Tablet) to
send a variety of student data and request easily through cloud
computing to obtain different decisions and levelling reports.
Each network has several application hosts = (H
1
, H
2
, H
n
)
providing the SaaS, and can be allocated to execute the cloud
stakeholders that make the intelligent decisions. Each
application host has a set of resources = (R
1
, R
2
and R
n
) that
can be allocated for the coming school requests. Each network
has a network administrator that is responsible for the
coordination of the communication between the hosts inside
the networks and other networks in the cloud. Network
administrator is responsible for running of the algorithm based
on the foundation shown in Fig. 2.
Fig. 2 Proposed system framework overview
A. Initial Suggested Dataflow on the Cloud and ML Process
Different algorithms have different outcome on the outcome
of the model and its performance. In order to select the
algorithm that performs the best from the start will be difficult
because the caret package in R has 237 models, and out of
those, 189 can possibly be used for classification problems,
[10], [12]-[14].
As shown in Fig. 2, each of the school will be having access
to the system and data moved to and from the cloud will be
stored there. The cloud is a virtual machine itself and then
further split into small virtual machines separated by each
port. Apache server being the first server to interact with the
user, it will act as the fog layer and then it is connected to the
database and finally the collected dataset will be passed onto
the ML tools. The data flow among the servers will be done on
a controlled basis through the web application to the apache
server and thereby to the database. The data in whole are
processed with ML algorithm which is supervised decision
making. The data are processed and analysed depending on
the results patterns and analysis. The logics were embedded
and pushed to the web application for each node which does
the task of filtering and predicting data.
We will process new datasets of students with parameters to
look for ML algorithm to give out results and hereby checking
for the results. Once the results have a satisfactory outcome as
per expectations, the data will be processed via TIBCO tools
to be presented back to the nodes which are authorized to
visualize the data. Keeping the ideology of student data, the
possibilities of results could be individual algorithms for the
levelling, prediction of performance, analysis of student’s
migration and many more data sets and patterns can be
visualized and studied. One perspective of levelling done by
the brain.js will be subjected to update depending on the ML
results. TIBCO tolls also further extend to the process of
reporting and analytical dashboard which shall be embedded
to the view of the nodes depending on the permission level
provided by the routes and controller.
World Academy of Science, Engineering and Technology
International Journal of Humanities and Social Sciences
Vol:14, No:12, 2020
1183International Scholarly and Scientific Research & Innovation 14(12) 2020 ISNI:0000000091950263
Open Science Index, Humanities and Social Sciences Vol:14, No:12, 2020 waset.org/Publication/10011632
B. System Data Modelling
There are many factors that have an influence on student
admissions, levelling and differentiation, as shown in Table I.
TABLE I
PRELIMINARY FACTORS THAT HAVE AN INFLUENCE ON STUDENT ADMISSION,
LEVELLING AND DIFFERENTIATION
No Quantity Quantity Description
1 Student ID Numerical 0001 +
2 First Name Nominal Given Names
3 Last Na me Nominal Family Name
4 Gender Nominal Male / Female
5 Nationality Nominal British, American, UAE, etc.
6 Previous School Name Nominal Belvedere British School,
GEMS American school, etc.
7 Previous Curriculum Nominal American, British, MOE, etc.
8 New School Name Nominal Belvedere British School,
GEMS American school, etc.
9 New Curriculum Nominal American, British, MOE, etc.
10 Previous year/grade Numerical FS1/PRE-K, FS2/KG1,
Year1/KG2, etc.
11 Proposed Year/grade Numerical FS1/PRE-K, FS2/KG1,
Year1/KG2, etc.
12 Interview Status Nominal Passed / Failed
13 Math Entry Exam Mark Numerical 0% - 100%
14 Science Entry Exam
Mark
Numerical 0% - 100%
15 English Entry Exam Mark Numerical 0% - 100%
16 End of Term 1 Math
Mark
Numerical 0% - 100%
17 End of Term 1 Math
Mark
Numerical 0% - 100%
18 End of Term 1 Math
Mark
Numerical 0% - 100%
19 End of Term 2 Science
Mark
Numerical 0% - 100%
20 End of Term 2 Science
Mark
Numerical 0% - 100%
21 End of Term 2 Science
Mark
Numerical 0% - 100%
22 End of Term 3 English
Mark
Numerical 0% - 100%
23 End of Term 3 English
Mark
Numerical 0% - 100%
24 End of Term 3 English
Mark
Numerical 0% - 100%
VI. CONCLUSION AND FUTURE WORK
Schools in multicultural countries are constantly levelling
students without the use of automation. Therefore there is a
need to implement the suggested framework in order to assist
schools in student transition between schools while also have
access to a unbiased grading system that can predict student
level based on previous data.
For the future work, there are additional researches to be
conducted in order to be updated about the current
development of our area of interest. We will continue on
developing the computational framework and implementing
the framework with the use of different tools to be able to
facilitate student admissions on other curricula. Further we
will design and develop the sophisticated ML software that
will generate the level of the student and provide
differentiation information for the teachers. To test the system
we will use data that we have already collected. Training of
the algorithm will be commenced once the system is tested to
be functioning as per expectations. As we have schools that
contributed to this research by providing us with valuable
data, we aim at implementing the system on those schools and
gather results and findings to compare them with the
simulation results to discover the reliability of the system.
REFERENCES
[1] ADEK, “No Title,” 2019. (Online). Available:
https://www.adek.gov.ae/. (Accessed: 17-Sep-2019).
[2] R. Al-Shabandar, A. Hussain, A. Laws, R. Keight, J. Lunn, and N. Radi,
“Machine learning approaches to predict learning outcomes in Massive
open online courses,” Proc. Int. Jt. Conf. Neural Networks, vol. 2017-
May, pp. 713–720, 2017.
[3] N. Sultan, “Cloud computing for education: A new dawn?,” Int. J. Inf.
Manage., vol. 30, no. 2, pp. 109–116, 2010.
[4] R. Almajalid, “A Survey on the Adoption of Cloud Computing in
Education Sector,” pp. 1–12, 2017.
[5] M. Britland, “No Title,” 2013. (Online). Available:
https://www.theguardian.com/teacher-network/teacher-
blog/2013/jun/19/technology-future-education-cloud-social-learning.
(Accessed: 20-May-2019).
[6] D. G. Chandra and D. Borah Malaya, “Role of cloud computing in
education,” 2012 Int. Conf. Comput. Electron. Electr. Technol. ICCEET
2012, pp. 832–836, 2012.
[7] F. Q. Khan, M. Ishaq, A. I. Khan, and B. Soubani, “Adapting Cloud
Computing in Higher Education,” vol. 5, no. 11, pp. 823–830, 2014.
[8] A. Rao, “Database as a Service in Cloud Computing,” Cc.Gatech.Edu,
vol. 7, no. 3, pp. 389–396, 2018.
[9] T. Hendrickx, B. Cule, P. Meysman, S. Naulaerts, K. Laukens, and B.
Goethals, “Mining association rules in graphs based on frequent
cohesive itemsets,” Lect. Notes Comput. Sci. (including Subser. Lect.
Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 9078, no. 3, pp.
637–648, 2015.
[10] C. E. Brodley, T. Lane, and T. M. Stough, “Knowledge discovery and
data mining,” Am. Sci., vol. 87, no. 1, pp. 54–61, 1999.
[11] C. Romero, S. Ventura, P. G. Espejo, and C. Hervás, “Data mining
algorithms to classify students,” Educ. Data Min. 2008 - 1st Int. Conf.
Educ. Data Mining, Proc., pp. 8–17, 2008.
[12] A. J. Stimpson and M. L. Cummings, “Assessing intervention timing in
computer-based education using machine learning algorithms,” IEEE
Access, vol. 2, pp. 78–87, 2014.
[13] M. Huber, C. Kurz, and R. Leidl, “Predicting patient-reported outcomes
following hip and knee replacement surgery using supervised machine
learning,” BMC Med. Inform. Decis. Mak., vol. 19, no. 1, pp. 1–13,
2019.
[14] Ł. Wiechetek, M. Mędrek, and J. Banaś, “Business Process Management
in Higher Education. The Case of Students of Logistics,” Probl. Zarz.,
vol. 15, no. 4 (71), pp. 146–164, 2018.
World Academy of Science, Engineering and Technology
International Journal of Humanities and Social Sciences
Vol:14, No:12, 2020
1184International Scholarly and Scientific Research & Innovation 14(12) 2020 ISNI:0000000091950263
Open Science Index, Humanities and Social Sciences Vol:14, No:12, 2020 waset.org/Publication/10011632
