IoT and AI Based Student’s Attendance Monitoring System

to Mitigate the Dropout in Non-boarding Secondary

Schools of Rwanda: A Case Study of Wisdom School
Musanze
Munyaneza Midas Adolphe, Gasana James Madson, Uwimana
Josephine, Shumbusho Jean Pierre, Nzayisenga Joselyne, Gafeza
Gaspard, and Niyonzima Martin
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

IoT and AI Based Student’s Attendance Monitoring System to

Mitigate the Dropout in Non-boarding Secondary Schools of
Rwanda: A Case Study of Wisdom School Musanze

1Munyaneza Midas Adolphe, Abstract

1Gasana James Madson,
1Uwimana Josephine,
Purpose: This project aimed to test an IoT and AI
1Shumbusho Jean Pierre,
based system that monitor students from home to
2Nzayisenga Joselyne, 2Gafeza
schools, during class hours and from school to home
and notify parents and school administrators about
Gaspard, and 3Niyonzima the irregularity observed to their respective children.
Martin
1 Methodology: In this project, secondary data was
Lecturer, Rwanda Polytechnic- used and was retrieved from the school’s record of
Integrated Polytechnic Regional Wisdom School Musanze located in Musanze
College Kigali (IPRC-KIGALI). District. The main data to consider were sex whether
2
Postgraduate Student, Regional male or female. Another important data was
Center of Excellence in orphanage,whether pupil is orphan or not orphan,
Biomedical Engineering and and school fees payment by checking whether
Health (CEBE) - University of student paid school fees or had not paid. These
Rwanda. mentioned data were taken randomly from senior
3 one (S1) to senior six (S6) in academic year 2020-
Instructor, Rwanda Polytechnic-
2021.
Integrated Polytechnic Regional
College Kigali (IPRC-KIGALI). Findings: The system is equipped of a finger print
sensor to register and verify students and staff
attendance, a Passive Infrared (PIR) sensor to detect
Article History the presence of human to wake-up the device, a real
Received: 14th Dec 2022 time clock to synchronize each generated report with
the local time. A web application is developed to
Received in Revised Form: 15th allow students real-time monitoring for parents and
Mar 2023 school administrators and the system is be able to
Accepted on: 21th Mar 2023 generate a daily, monthly and annually report.
Unique contribution to theory, practice and
policy: Classification machine learning with
decision-tree algorithm is used to analyze data and
generate a model to evaluate the impact of
monitoring attendance on preventing students to
dropout. The generated model with accuracy of
91.4% shows that keeping students’ attendance at
high percentage would reduce significantly the
dropout rate in secondary schools of Rwanda.
Keywords: Biometric, internet of things, sensor,
fingerprint, regression and decision-tree.

27
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

INTRODUCTION
Personal identification is an important aspect in recognizing the identity of a particular
individual. A person’s identity is validated through the traditional or biometric methods;
biometric technology associated with Internet-of-things (IoT) provides a framework for
integrated computing devices capable of managing efficiently the students’ attendance.
According to Jain et al. (2000), there are two types of traditional methods which are token-
based and knowledge based identifications. Examples of the token-based method include
possession of a passport, driving license, and different types of cards such as identity (ID) card
and credit card. Although it is convenient to carry these identity documents, these documents
can be reproduced, stolen, or lost. On the other hand, the knowledge-based method is related
to a password or personal identification number (PIN) created by each individual for
authentication. Nonetheless, it tends to be forgotten easily, especially if the person has several
passwords or PINs for different applications. Another alternative method is through biometric
adoption, which considers the physical or behavioral characteristics distinctive to an individual.
Physical characteristics refer to inherent features of the human body part. These include the
face, fingerprints, and iris. On the contrary, behavioral characteristics deal with features
observed from human action. Examples of human action are gait, voice, and signature (Dinca,
& Hancke, 2017).
According to German and Barber (2017) application of biometrics in different domains.
Biometric-based attendance recording or tracking system is part of the low percentage of
biometric applications in the education sector, Attendance marking is considered as one of the
crucial parts of a class. This is to ensure students participate in the class activities and learn
from their educators. Student academic attendance is very crucial on account that it's going to
have an effect on students from gaining knowledge, skills and their grades.
Currently in Rwanda, most of the school attendances are marked using the conventional
methods such as calling out names or signing off attendance on paper. Unfortunately, these
methods are not suitable for a large class. It is a waste of time for the educator to call out all
the students’ name to mark their attendance whereby the time can be utilized more efficiently
for the teaching and learning processes. Another method is by signing off attendance.
Distraction happens because the student needs to sign the attendance list and pass on to the
other student during class.
Moreover, it can be compromised by a student who signs on behalf of their friend who does
not attend the class. Therefore, biometric systems have been developed as an alternative way
to mark attendance in class. One of the advantages is that the student cannot manipulate the
attendance as each individual has different biometric characteristics. In addition, it also helps
in improving efficiency and reduces the educator’s burden as the attendance is marked
automatically. This project is about taking the student fingerprint templates and matching them
to fingerprint stored in database to confirm their attendance. The predominant purpose of
carrying out this project is to design and implement an IoT-based student attendance system
where the data are collected from a remote fingerprint device and sent to a web-based
application for analysis and reports are automatically generated to concerned educational
partners.
Problem Statement
The Government of Rwanda set policies to improve school progression and completion rate to
reduce the numbers of students dropping out in secondary school but the dropout rate is still
very high especially in non-boarding schools. The students leave their home early morning and

28
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

go to school and return back in evening, through the whole day there is no student’s monitoring
system for parents and school managers to ensure that students really attended classes. For
some students, instead of attending classes they go in illegal practices. It is in this background
that this research is contributing to reduce the dropout by developing a real time IoT and AI
based student’s attendance monitoring system that allow communication between parents and
school managers, through SMS or Email notification.
Rwanda Education Statistics Related to Dropout
Rwanda has established a 12 years’ basic education that include 6 years of primary school and
6 years of secondary school; most of these secondary schools are non-boarding school.
Table 1: Lower secondary school student’s drop out from 2013 to 2019

Source: (MINEDUC, 2020)

Table 2: Upper secondary school student’s drop out from 2013 to 2019

Source: (MINEDUC, 2020)

LITERATURE REVIEW
Research was carried out to identify the requirements of any biometric system, and specifically
of a fingerprint system for use within the education environment, before comparing fingerprint
suitability with that of other biometric techniques. A choice of fingerprint device had to be
considered, and consideration given to the role of identification or verification, so that the best
method of operation could be applied. Some biometrics attendance in different regions and
show their drawbacks in education especially in case of Rwanda.
Tiwari et al. (2015) In designing a biometric based application seven characteristics have been
identified for consideration in any system, which uses biometrics for authentication of
individuals and the comparison of fingerprint techniques with the other biometrics methods is
given here. This comparison is completely based on the scientific literature but examples are
given in an attempt to elucidate the findings. The new school surveillance technology doesn’t
turn off when the school day is over: anything students’ type in official school email accounts,
chats or documents is monitored 24 hours a day, whether students are in their classrooms or
their bedrooms. Tech companies are also working with schools to monitor students’ web
searches and internet usage, and, in some cases, to track what they are writing on public social
media accounts.
In (Freebrowsinglink, 2021) Chinese schools are using ‘smart uniforms’ to track their students’
locations. The reason of tracking the children is to encourage better attendance rates. For
29
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

tracking, each of student uniform has two inbuilt chips in the shoulders. This enables the
appropriate management by knowing when a child exists/leaves the school premises. Face
Identification is built right inside the uniforms and there is a small quantity of facial recognition
software at the entrances to make sure that the right student is wearing the right outfit. There
is Alarm system to keep the school premises in order, where by when student try to leave during
school hours, an alarm goes on and there is change if he/she is allowed to go or not. The chips
can apparently detect when a student has fallen asleep in class, and allow students to make
payments (using additional facial or fingerprint recognition to confirm the purchase).
Despite having those little technologies built right inside those clothing, students can also wash,
and rewash their uniforms without any hindrances. The developers on this smart uniforms are
Guizhou Guanyu Technology Company, confirmed that each outfit can endure up to 150
degrees Celsius (302 degrees Fahrenheit) and 500 times of washes. Luminositylab (2021)
Guardian Drones is an autonomous drone system designed by Luminosity researchers to
increase university campus safety by augmenting existing emergency response teams with
autonomous agents. Using a proprietary mobile application, students can request to be escorted
through campus by an autonomous drone. The drone fleet provides illumination and establishes
a live video link for ASU’s Police Department to monitor activity. To ensure the success of the
system, the drones were designed with multiple functional safety features, such as deployable
parachutes and encryption protected software. The drone fleet is capable of responding to a
student’s request and arriving at their location on campus within a minute. The system was
designed to be self-sufficient, in addition to the development of the drones, Luminosity students
developed and submitted an invention disclosure for a base station that is capable of
automatically docking and charging drones through a contact driven system.
Narzullaev and Narzullaev (2021) proposed a new AI-based system that can automatically
detect the student’s presence or absence in a classroom using the Wi-Fi signal information
collected from student’s smartphone. The proposed system does not require any additional
investment in hardware or infrastructure and achieves up to 94% accuracy by implementing
the Logistic regression-based machine learning classification algorithm. This system is not
applicable in Rwanda because the rule of ministry of education (MINEDUC, 2020) doesn’t
permit the use of smartphones in secondary schools.
Shoewu et al. (2011) proposed an embedded computer-based lecture attendance management
system where a single-chip computer based subsystems (an improvised electronic card and the
card reader) were interfaced serially to the serial port of the digital computer. The electronic
card is a model of a smart card containing the student identity (ID Name, Matriculation Number
and five pin encrypted code). The student ID is authenticated by the card reader which
compares the entrance code with the encrypted code on the card swiped through the card reader.
The student is granted and or denies specific lecture attendance based on the result of the
comparison by the backend software system running on the PC to which the card reader is
serially interfaced. The system though provided a simplified, low cost embedded computer
based system solution to the management of lecture attendance problem in developing
countries but does not eliminate the risk of impersonation. The system is devise-based in which
students have to carry RFID cards and also the RFID detectors are needed to be installed.
Shehu and Dika (2010) proposed a real time computer vision algorithms in automatic
attendance management systems using Computer vision and face recognition algorithms and
integrating both into the process of attendance management. The system eliminates classical
student identification such as calling student names, or checking respective identification cards,
but still lacks the ability to identify each student present in class thereby providing a lower

30
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

recognition rate because facial images are subject to change between the time of enrolment and
time of verification and also poses a bigger financial burden during installation and does not
offer any privacy protection. Jomon and Zacharia (2013) proposed a face recognition based
attendance system based on Eigen face recognition. Images are converted into Eigen faces,
Recognition is performed by comparing Eigen face got from the input image and Eigen faces
in the database. The problem with this approach is that this method is very sensitive to face
background, head orientations and it doesn’t recognize the face of a person if the person is
wearing glasses or a grown beard, etc.
There are also papers involving using biometrics for attendance (Shoewu & Idowu, 2012)
fingerprint is used for marking the attendance of students. There is a biometric sensor which
takes the fingerprint, feature extraction done on that data. If it is for enrolment, then that data
is stored in the database else if it is for authentication then that data is started matching with
the data in the database. The problem with this method is that for attendance students should
go to the place where this hardware device is located or pass the hardware device around the
students during class which can be a distraction to the students. (Chaniago & Junaidi,2016)
presenting a student attendance application based on the SMS Gateway application that was
made to ensure that students attend classes, with the following steps: students must submit an
identity card to the teacher in charge, then the teacher scans the barcode printed on the card,
and finally the data is stored in the database automatically. This application provides services
for parents to find information on a weekly and monthly basis. This application can also be
taken into consideration in making decisions for principals by downloading student attendance
data in the Microsoft Excel file format. The implementation of SMS gateway-based
applications for student attendance in schools is expected to be the solution to students'
absenteeism problems.
RESEARCH METHODOLOGY
Methodology refers to the procedures or techniques used to collect data and analyse them. They
are many techniques and this project used data collection, data analysis to generate a model
then prediction model whose outcome is tp predict dropout and future use of information
extracted from provided data.
Data Collection
In this project, we collected primary data from the school by installing a device and then
students and staff need to enroll to the system at the beginning and then their attendance is
taken and stored in database and data captured sent on cloud platform. With the time limitation,
we decided to use secondary data taken manually by teachers and school’s administrators, the
used data are obtained from the school’s record of Wisdom school Musanze located in Musanze
district. These are from six classes taken randomly from senior one (S1) to senior six (S6) in
academic year 2020-2021.
After interview with school staff, parents, students and education partners, it is realized that
there is no available record of student’s implication in drug abuse or teenage pregnancy but
found that dropping out from the school keeps existing, This project helped much since it
monitor and provide data and predict the dropout and this can help to mitigate the right solution
to solve it by basing to the data and facts. A field of data was included in this research related
to the students punished for escaping school during class hours because they are suspected to
go in illegal activities such as drug consummation and adultery.
According to the literature some researchers shown that student’s dropout might be influenced
by parent poverty, being orphan or age, the database includes such data related to payment

31
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

historic, which is in charge of a student and their age. Finally, we included a field of data related
to attendance (in percentage) because I assuming that by daily monitoring attendance I can
reduce the case of drug abuse and teenage pregnancy, hence reduce the dropout. In total 240
students’ data were gathered from S1 to S6. The data base contains the following: students
‘ID, class, age, orphan, pay school fees, attendance and dropout.
Data Analysis
This is taken as a tool used to extract useful information from raw data collected and stored in
database to help to make decisions. Collected data are accessed by a cloud platform that
generates insight information to education partners. Multiple algorithm of machine learning
were applied to the data to generate a model with high accuracy which shows the contribution
of different variable to the school dropout.
Data Interpretation/Reporting Plan
Information gathered are shared to the end users like parents, school managers and education
decision makers.
SYSTEM DESIGN ANALYSIS AND RESULTS
System Design
The student’s attendance monitoring system consists of three main parts. The first is IoT
hardware component that build the entire circuit. This part involves sensing components, data
processing component and actuators components. The second part consists of programming to
make system hardware operational. The last part is data processing and analytics, which is done
using Python programming and regression as machine leaning technique and it is visualized on
graph.
System Design Block and Circuit Diagram
The system is equipped of a finger print sensor to register and verify students and staff
attendance, a PIR sensor to detect the presence of human to wake-up the device, a real time
clock to synchronize each generated report with the local time.

Figure 2: Block diagram of the attendance monitoring system

Solar energy harvesting is used to charge the device’s internal battery, the system requires a
router to connect to the internet and a web application is developed to analyse data and generate
insight information. A web application also allows students real-time monitoring for parents
and school administrators and the system is be able to generate a daily, monthly and annually
report to education decision makers.

32
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

Circuit Description
The first thing is to do in the program is to include all the required libraries. Here in my case, I
have included “Adafruit_Fingerprint.h” for using R305 fingerprint sensor and
“ESP8266WiFi.h” for using ESP8266 NodeMCU Wi-Fi module. To use the I2C interface of
an LCD display I used LiquidCrystal_I2C.h library. Then we configured the serial port in
which fingerprint scanner is connected. In my case, I declared D5 as RX Pin and D6 as TX pin.

Figure 3: Project’s circuit diagram

Next, I set my Wi-Fi router credentials in the program. Replace your network SSID name in
place of “xyz” and password in place of “123456789”. As we are going to use Pushing box
API for sending the data to Google sheet, then I have to check for fingerprint sensor
connectivity. This confirmed us regarding the fingerprint sensor’s successful pairing with the
NodeMCU. If the connection to the server is successfully established, then a complete URL is
created using the device ID which is got in pushing box API. If the client doesn’t respond for
more than 5 seconds, it shows client timeout. Otherwise, it sends the data to Google sheet using
the URL via pushing box API.

33
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

System Design Flowchart

Figure 4: Flowchart of the attendance system

A flowchart is a graphical or symbolic representation of a process so that the process is more
likely understood and applied correctly and consistently and it is easy to communicate it to
other people. The above flowchart shows the steps of process undertaken by the attendance
monitoring system when a teacher has to get the students ‘attendance in his respective course.

34
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

System High-level Design

Figure 5: System high-level design

System Analysis
Algorithm to use
An algorithm is said to be a procedure or formula used to solve a problem.so we used codes to
perform manipulations to achieve the results. More explanations and flow chart are given in as
we proceed. Due to the type of output to predict which is categorical instead of numerical
values, we found that classification algorithm is best fitting among others. Classification is the
type of supervised machine learning which is mostly fitting in a predictive modeling problem
where class or labels are categorical instead of numerical values. I chose classification among
others because we were predicting labels like dropout or no dropout.
Machine learning is a method of data analysis that automates analytical model building based
on the idea that systems can learn from data, identify patterns and make decisions with minimal
human intervention. Supervised learning is one of the most basic types of machine learning
among others like unsupervised learning and reinforcement learning; here we focused on
supervised learning because of the nature of the data and predicted output. The dataset obtained
from the attendance system are saved in excel sheet and was used to train model.

35
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

Table 4: Attendance monitoring system dataset

The dataset shows that there are 240 students in 6 classes and among them at least 40 students
were dropped out from school.

Figure 6: Non-dropout vs dropout

There is a problem of class imbalance between dropout and No_ dropout; hence it is needed to
perform oversampling to fix this problem.

Figure 7: fix the problem of class imbalance

Some variables need to be pre-processed to be analyzed

Class input variable: In order to reduce the number of coefficient I decided to use lower
secondary (0) and upper secondary (1)
36
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

Sex input variable has two fields, female (F) and male (G)
Orphan input variable: has two fields, orphan (Y) and not orphan (N)
Pay school fees: two fields pay school fees (Y) and not pay school fees (N)
Punished for escaping class during class ours: two field punished (Y) and not punished (N)
Table 5: Pre-processed dataset

Three models of classification machine learning; Logistic Regression, K-Nearest neighbors and
Decision Tree have been applied to the data to see which one is more accurate. Here are the
results. For Logistic regression an accuracy of 81.9 % was obtained as shown in table 6.
Table 6: Logistic regression result

For K-Nearest Neighbors an accuracy of 81.9 % was obtained as detailed in table 7.

Table 7: KNN result

Finally, for Decision Tree an accuracy of 91.4 % was obtained.

37
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

Table 8: Decision three result

RESULTS DISCUSSION
Impact of sex on dropout
The analyzed data shows that the number of girls (F) is almost equals to the number of boys
(G), but more than 25% of girls are dropping out while on boys’ dropout rate is about 18%,
this is due to the fact that every student who get pregnant end up by dropping out.

Figure 8: Impact of sex on dropout

Students’ dropout vs secondary school levels
The data shows that the dropout rate is high in lower secondary school and it is reduced as the
students’ progress toward upper secondary school. There are about 9 dropouts over 43 students
in senior one (187%) and this rate becomes 5 over 30 students with a percentage of 16.6%.

Figure 9: Students’ dropout vs levels

38
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

Students’ dropout vs parent’s poverty

The parents’ poverty which was indicated by means of school fees payment either early or
very late. In a total of 240 students, 155 of them got school fees on time(at the time of starting
school) and 20 students dropout (12.9%), while 95 paid school fees with difficult among them
20 students dropout (20%). This showed that the parents’ poverty is one of the factors that
influence the dropout.

Figure 10: Students’ dropout vs parents’ poverty

Students’ dropout vs student’s punishment
Among 145 students who do not punished for escaping class during class hours, only 10
dropped out (6.9%) while among 95 students who were punished due to escaping classes 30
30 of them dropout which is 31.5%. This indicated that punishments due to class escape
contribute to student’s dropout.

Figure 11: Students’ dropout vs students’ punishment

39
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

Students’ dropout vs students’ age and attendance

Figure 12: Students’ dropout vs students’ age and attendance

The dataset analyzed shows that students in secondary school belong to the range from 12 years
to 20 years old, the figure 12 indicate that the age does not contribute to the dropout as the
dropout cases are distributed to the whole range of age. In opposite, the attendance contributes
a lot to the dropout as the students with greater than 70% of attendance do not dropout. So, to
monitor the students’ attendance would reduce significantly the dropout rate in secondary
schools.
Results discussion from the model
In this analysis 75% of the dataset have been used for model training and 25% were used for
model testing, the algorithms are used to train data and generate prediction results using three
different models’ algorithm and the results are obtained with different accuracies: Decision
Tree had an accuracy of 91.4 %, K-nearest neighbors had an accuracy of 81.9 % and logistic
regression we have an accuracy of 81.9 %. It is clear that classification with decision tree is
generating a model which mostly feet to the collected data to predict the dropout in secondary
school.
Table 9: Results comparison from three models

SN Prediction model Dropout Precision Recall f1-score Support Accuracy

1 Logistic yes 0.82 0.8 0.81 51 81.90%
Regression non 0.82 0.83 0.83 54
macro avg 0.82 0.82 0.82 105
weighted avg 0.82 0.82 0.82 105
2 K-Nearest yes 0.76 0.76 0.76 51 81.90%
Neighbors non 0.78 0.78 0.78 54
macro avg 0.77 0.77 0.77 105
weighted avg 0.77 0.77 0.77 105
3 Decision Tree yes 0.85 0.78 0.82 51 91.40%
non 0.81 0.87 0.84 54
macro avg 0.82 0.83 0.83 105
weighted avg 0.82 0.83 0.83 105

40
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

It is true that some other factors contribute to dropout like being orphan and social economic
situation of the students’ family but those factors can’t be controlled, the attendance on contrary
when monitored students were somehow forced to attend even if in some situation they could
not without such mechanism. It is clear that decision Tree provide a model with high accuracy
and the dropout is given by:
Dropout = 0.623Female_sex + 0.447Male_sex + 0.909No_orphan + 0.161Orphan +
0.861No_fees + 0.209Fees + 0.207Not_punished+ 0.863 Punished, + 0.099 Age -
0.078Attendance +1.113Lower_secondary -0.042Upper_secondary + 1.07
As the attendance has a negative coefficient (-0.078) this shows that attendance is indirectly
proportional to the students’ dropout and when the attendance increases the dropout decreases
considerably so monitoring the attendance to keep it to high percentage reduces the dropout
rate.
CONCLUSION
This project achieved its main purpose which was to design and implement an IoT based
attendance monitoring system that monitor non-boarding secondary school’s students from
home to schools, during class hours and from school to home and notify parents and school
administrators about the irregularity observed to their respective children in order to reduce the
student’s dropout rate. A fingerprint-based device is implemented for students and staff
identification, authentication and evaluation while making attendance. The system is
generating a real time report to parents and education decision maker and provides an efficient
co-parenting communication regarding student’s behaviors during class hours.
In this project a model that shows the impact of monitored attendance on preventing students’
drop out in secondary schools is generated with an accuracy of 91.4% and it confirm the
hypothesis because shows that keeping the attendance at high percentage would reduce
significantly the students’ dropout rate, but the model is not deployed to the system.
RECOMMENDATION
During the implementation of the project it was realized that students of Embedded Computing
System were missing a lot of knowledge of programming especially on web application &
database development and data analysis algorithm development.it is recommended to the
center to find how to include such modules to fill the gap.
The system uses a single biometric technology (fingerprint) which has some drawbacks when
there is skin degradation, so it is recommended to the future researcher to combine more than
one technology to implement a more efficient recognition system.
The data set used in this project was taken from a single school and a small sample of 240
students data of a single academic year were analyzed; this might not output excellent results
of students’ dropout. Hence, it is recommended to the future researcher to expand sample to
different schools of different districts and analyses data of more than one academic year.
The study recommend that further studies may be applied to this project where a model will be
deployed to the system and shows the student’s dropout probability based on system’s database
information.

41
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

REFERENCES
Chaniago, M. B., & Junaidi, A. (2016,). SMS Gateway and barcode technology for presence
of students in SMK Unggulan Terpadu PGII Bandung: A case study. In 2016 4th
International Conference on Cyber and IT Service Management (pp. 1-4). IEEE.
Dimitri et al., (2018). Assessment of Repetition and Dropout in Basic Education in Rwanda
Full Report From data to policy Acknowledgements.Rwanda,kigali
Dinca, L. M., & Hancke, G. (2017, June). Behavioural sensor data as randomness source for
IoT devices. In 2017 IEEE 26th International Symposium on Industrial Electronics
(ISIE) (pp. 2038-2043). IEEE.
Freebrowsinglink. (2021). https://www.freebrowsinglink.com/chinese-schools-using-smart-
uniforms/.
German, R., & Barber, K. S. (2017). Current biometric adoption and trends. The University of
Texas at Austin. Retrieved from identity. utexas.
edu/assets/uploads/publications/Current-Biometric-Adoption-and-Trends. pdf.
Jain, A., Hong, L., & Pankanti, S. (2000). Biometric identification. Communications of the
ACM, 43(2), 90-98.
Jomon Joseph1, K. P. Zacharia. (2013). Automatic Attendance Management System Using
Face Recognition. International Journal of Science and Research (IJSR).
Kanyoni, M., Gishoma, D., & Ndahindwa, V. (2015). Prevalence of psychoactive substance
use among youth in Rwanda. BMC research notes, 8(1), 1-8.
Luminositylab. (2021). https://theluminositylab.com/projects/guardian-drones/.
MINEDUC. (2020). Rwanda education statistics 2018. kigali. Retrieved 10 15, 2021, from
http://197.243.16.104/~mineduc/newweb/fileadmin/user_upload/pdf_files/2018_Rwa
nda_Education_Statistics.pdf
Narzullaev, A., Muminov, Z., & Narzullaev, M. (2021, July). Wi-Fi based student attendance
recording system using logistic regression classification algorithm. In AIP Conference
Proceedings (Vol. 2365, No. 1, p. 030004). AIP Publishing LLC.
Ngamije, J., & Yadufashije, C. (2016). Drugs use among the youth In Rwanda: A public
health concern and family’challenge. International Journal of emergencies computer
and technology, 3(9), 4591-4603.
Seifedine, K., & Mohamad, S. (2010). Wireless attendance management system based on iris
recognition. Scientific Research and essays, 5(12), 1428-1435.
Shehu, V., & Dika, A. (2010, June). Using real time computer vision algorithms in automatic
attendance management systems. In Proceedings of the ITI 2010, 32nd international
conference on information technology interfaces (pp. 397-402). IEEE.
Shoewu, O., Olaniyi, O. M., & Lawson, A. (2011). Embedded Computer Computer-Based
Lecture Attendance Based Management System.
Shoewu, O., & Idowu, O. A. (2012). Development of attendance management system using
biometrics. The Pacific Journal of Science and Technology, 13(1), 300-307.

42
European Journal of Technology
ISSN 2520-0712 (online)
Vol.7, Issue 1, pp 27 - 43, 2023 www.ajpojournals.org

Srinidhi, M. B., & Roy, R. (2015, January). A web enabled secured system for attendance
monitoring and real time location tracking using Biometric and Radio Frequency
Identification (RFID) technology. In 2015 International Conference on Computer
Communication and Informatics (ICCCI) (pp. 1-5). IEEE.
Tiwari, T., Tiwari, T., & Tiwari, S. BIOMETRICS BASED USER AUTHENTICATION.
Uwizeye, D., Muhayiteto, R., Kantarama, E., Wiehler, S., & Murangwa, Y. (2020).
Prevalence of teenage pregnancy and the associated contextual correlates in
Rwanda. Heliyon, 6(10), e05037

43