![International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
@ IJTSRD | Unique Paper ID – IJTSRD43700 | Volume – 5 | Issue – 4 | May-June 2021 Page 1550
introducing this problem and finding the root cause of the
increased dropout rates. The findings have been effective in
understanding that the prediction of a dropout rate if
determine for a student it can be effectively useful in
addressing the problems and the concerns of the student.
This information is valuable in reducing the rate ofdropouts
for the students and improving the rate of successful
completion for the university. The prediction has also been
useful in understanding the inconsistencies in the course
material which can be rectified in time before a lot of
student’s dropout. Therefore the related words analyzed in
this approach have been effective in the designing of our
technique for the purpose of dropout predictionthroughthe
implementation of machine learning approaches. Literature
survey paper provides a thorough analysis of the student
dropout rate prediction and realizes an effective machine
learning implementation for improving the dropout
prediction.
In this paper, section 2 is dedicated for the literature review
of past work, Section 3 describes the details of the
developmental procedure of the model. Section 4 evaluates
the results through some experiments and finally section 5
concludes this research article with the traces of the future
scope.
Literature Review
T. Hasbun narrates in Organization for Economic Co-
operation and Development countries student dropout is
one of the salient problems. [1] Thus for dropout prediction,
Educational Data Mining (EDM) analysts are studying
institutional interventionsthroughstudents.Intheproposed
paper the authors have given importance to extracurricular
activities from science degrees engineering and business to
predict dropout. The main aim of this paper is to prove that
how extracurricular activities can be suitable factors for
dropout prediction. The results of the proposed technique
have shown better performance.
Di Sun states in the last few years there has been a rapid
growth in open online courses(MOOCs)butthey'rea serious
problem is arising by the MOOC researchersandprovidersis
dropout prediction. [2] The paper presents a solution for
students getting dropouts just by analyzing how much
syllabus can be completed by the students in a course. Thus
for predicting the student dropoutrecurrentneural network
(RNN) and learning resource representation layer is used
thus the accuracy of the proposed papers is higher than the
traditional machine learning methodology.
J. Canas aims to offer an easy and explainable plan of action
to recognize dropout-prone and fail-prone students.[3] For
practitioners and researchers studentswhodonotcomplete
there higher education or they interrupt their study in
between is one of the major concern. The early prediction of
the student’s dropout may help teachers to reduce dropout
rates and also suggest new learning materials for the
students from failing or not completing the course. In the
proposed paper the researcher has used tree-based
classification models to predict dropout-prone and fail-
prone students.
M. Mustafa describes for developing countries the national
and international loss of student’s dropout prediction is one
of the major concerns. The main aim of this paper is to
implement an effective dropout prediction system for
institutes and colleges.[4] To classify the successful from
unsuccessful students they implemented square test
methodology on factorssuchasfinancial condition,dropping
year, and gender. They have used feature selection, logistic
regression, and neural networks to data using data mining
techniques Classification and Regression Tree (CART) and
CHAm tree.
L. Wang explains the emerging wave of artificial intelligence
and network information platform technology for the
Massive Open Online Courses (MOOC). MOOC is a new
classroom model that is greatly hit on the traditional
education model.[5] In the proposed paper they
implemented a prediction model by using a time-controlled
Long Short-TermMemoryneural network (ELSTM).[5]Tasks
such as learning evaluation, talent training, curriculum
development, teaching mode have greater benefits. Some
many shortcomings and problems need to be solved in
MOOC. The experimental results of the proposed model are
higher which shows the effectiveness of theproposedpaper.
N. Shiratori states since the 1990s a serious problem has
been are rising in Japan that is dropout ratehasbeena major
point of interest by Yomiuri Shimbun. [6] As the semester
progresses students lose the interest in course. In
preliminary dropout, status distributes the dropout-related
state of students by using the logisticregressionmodel.Thus
the proposed paper implements model to divide students to
examine the issue of student dropout rates. Thus the
proposed research students with low motivation in studies
so faculty member to know the reasons for this low
motivation an early interview.
D. Ktoridou narrates that almosthalfofengineeringstudents
change their field due to teaching and advising insufficiency,
the complexity of the engineering curriculum.[7] There are
six personality types enterprising, conventional, artistic,
social, and realistic.Theproposedpapershowsthesimilarity
between engineering undergraduate students with
personality types. This research is done with 113
undergraduate students the main reason for leaving
engineering was poor teaching and advising. The paper
provides positive effects on both students and Engineering
departments
R. Pereira presents the student dropout prediction thus by
using the data mining techniques disciplinary and
institutional data of students from the University of Nariño.
[8] By using the classification techniques such as decision
trees Socioeconomic and academic student dropout profiles
were generated. The proposed model can help to develop
policies and strategies for the students to improve the
activities for studies. In recent years there have been many
types of research made in the field of education by using
Data mining techniques.
A. Ortigosa states the main aim of the proposed paper is to
put a stop to students from leaving the university.[9] The
proposed paper collected 11,000students’data fromthefive
years by using the C5.0 algorithm which is one of the
successful predictive models and SPA i.e Sistema de
Prediccioen de Abandono dropout prediction system in
Spanish. The SPA model was taken into use in 2017 and is
currently in use. Thus by using this model 117,000 risk
scores to detect the dropout risk of morethan5700students
it to support student dropout risk prevention.
K. Murakami performs machine learning techniques to
predict dropout rates of current and graduate students by
collecting their data thus the main aim of the paper is to](https://image.slidesharecdn.com/273discoveringstudentdropoutpredictionthroughdeeplearning-210721110815/85/Discovering-Student-Dropout-Prediction-through-Deep-Learning-2-320.jpg)
![International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
@ IJTSRD | Unique Paper ID – IJTSRD43700 | Volume – 5 | Issue – 4 | May-June 2021 Page 1551
dropout prevention. [10] The university arranges and
collects data of students toprovide enrollmentmanagement.
They focus on the student dropout and by using data from
both current, graduate, and dropout students. Thus byusing
the machine learning algorithm while predicting these
dropout rates it is important to investigate unknown items
regarding EMIR.
K. Santos explains for university student’s dropout
prevention by using various educational data mining
techniques. [11] The algorithm such as RF and Decision
Trees some algorithms are not successful to get high
outcomes due to the large data and a parameter tuning step.
Thus the proposed paper uses the different supervised
learning techniques such as Decision Trees, K-nearest
Neighbor, Support Vector Machines, Naive Bayes, Neural
Networks, and Random Forests to calculate the prediction
models for reducing school dropout.
R. Lottering describes in the field of education in South
Africa an advanced development has been seen in this
development dropout rate is noticed. [12] As in higher
education in South Africa student numbers grew by 32.8%
from 2006 to 2015 and the dropout rate of 17.1% has been
noticed and it can be doubled in the next ten years. Thus
proposed systems use data mining techniques and predict
dropout goals. The researches of the proposed papers show
a positive approach to the proposed model.
PROPOSED METHODOLOGY
Figure 1: Overview of the Proposed Idea
The proposed methodology of the learning outcomes of
massive online courses is depicted in figure 1 and it is
explained through following steps
Step 1: Preprocessing and Labeling – The proposed system
uses the dataset of massive online courses forthepurpose of
experimentation. The dataset is downloaded from the URL :
https://dataverse.harvard.edu/dataset.xhtml?persistentId=
doi:10.7910/DVN/29779.
Once Testing dataset of the user is received, it is statically
stored for the instance learning outcomes. And then
proposed model uses the Training set where it reads and
stored in a double dimension list, which is subjected to
preprocessing thereafter. This double dimension list of
dataset belongs to training data is divided into certified and
non-certified set based on the Boolean value present in the
column of the dataset.
Then this dataset is used to label them with the unique
integer numerical value using hash set functions. After this
all the required rows of the preprocessed list are assigned
with the integer numerical value which is then subjected to
clustering process in the next step of the model.
Step 2: K Nearest Neighbor – The preprocessed list attained
in the previous step is utilized as an input in this step of the
procedure. This step of the student dropout detection
performs the K Nearest Neighbor clustering. This is
performed to classify the dataset given as an input into
semantic groups to achieve accuratedetectionofthestudent
dropout in massive online courses. This procedure is
performed through the steps given below.
Distance Evaluation – The Euclidean Distance formula given
in the equation 1 below is being used for the purpose of
evaluation of the distance between the training and testing
dataset attributes. The dataset is provided in the form of a
list that is given as an input to this step of the procedure. The
data stored in the rows of this list is used for the purpose of
calculating the distance from the labeled testingdataset,and
stored as the row distance at the end of the list. This is
performed for all the entries in the list iteratively.
________ (1)
Where,
ED=Euclidian Distance
ATi=Attribute at index i
ATj= Attribute at index j
Centroid Estimation – The list with the data and the
respective row distances calculated in the previous step are
used in this step as an input. The list is sorted in the
ascending order of the Row distance and subjected to
random data point selection. These selected data points are
K in number and achieving the centroid. These data points
are used to determine the row distance of theselectedindex.
The obtained row distance is used further to achieve the
boundary of the clusters.
The row distances achieved previously are used to form the
average row distance of the entire list. These average row
distances along with the extracted row distance of the
centroid are used to determine the boundaries of the
clusters in the cluster formation procedure given below.
Cluster Formation – The centroids and the average row
distance acquired previously are providedasaninputtothis
step for cluster formation. This step then calculates the
boundaries of the clusters by addition and subtractionofthe
average row distance and the centroidrowdistancetoattain
the maximum and minimum values respectively. The data is
then subjected to these boundaries to form the clusters
which are aggregated into a cluster list. This cluster list
sorted into a descending order and the top 3 clusters are
provided to the next step as an input.
Step 3: Fuzzy Artificial Neural Network – This is one of the
most important steps in the proposed methodologywherein
a double dimensional list is utilized to effectively store the
attributes of the student details as well as the clusters of
these values. The fuzzy classification is achieved on this list
to effectively segregate the clusters and effectively identify
the difference between the maximum and the minimum
cluster values. The difference achieved is then effectively
divided into 5 equal parts to achieve effective fuzzy
classification labels. These labels correspond to the fuzzy
crisp values that are segregated as very high medium-low
and very low. The student ID and the score sum of different
activities are stored in a double list to generate the fuzzy
crisp ranges as mentioned in the below algorithm1.](https://image.slidesharecdn.com/273discoveringstudentdropoutpredictionthroughdeeplearning-210721110815/85/Discovering-Student-Dropout-Prediction-through-Deep-Learning-3-320.jpg)
![International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
@ IJTSRD | Unique Paper ID – IJTSRD43700 | Volume – 5 | Issue – 4 | May-June 2021 Page 1552
________________________________________________
ALGORITHM 1: Fuzzy Crisp Value generation
__________________________________________________
//Input : Student Score list SLST
//Output: FCLST ( Fuzzy Crisp List)
fuuzyCrispValueGeneration(SLST)
1: Start
2: Set min=0.5, max=0.5
3: for i=0 to size of SLST
4: TMPLST = SLST[i] [ TMPLST = Temporary Set]
5: PRSCRE= TMPLST [1] [ Score]
6: IF (PRSCRE < min)
7: min= PRSCRE
8: IF(PRSCRE >max)
9: max= PRSCRE
10: end for
11: RANGE1=0 , RANGE2=0 ,RLIST= ∅ [Rule List]
12: DF=( max-min)/5 [ DF= Diffrence Distance ]
13: for i=0 to 5
14: RANGE1=min
15: RANGE2=RANGE1+DF
16: min=RANGE2
17: TLST[0]= RANGE1 [TLST= Temporary List]
18: TLST[1]= RANGE2
19: FCLST = FCLST + TLST
20: end for
21: return FCLST
22: Stop
_________________________________
The neurons for the artificial neural network are then
classified according to the given rules above depending on
the classified certified set with labels provided such as very
high, high, medium, low and very low. These neurons are
then subjected to affective evaluationofthehiddenlayerand
the resultant probability achievedeffectivelydeterminesthe
dropout rate for the desired level according to the user
Step 4: Decision Tree – The probability values achieved from
the previous step of fuzzy artificial neural networks is taken
as an input to this step of the methodology.TheDecisiontree
approach is one of the most powerful and highly accurate
classification algorithms that completely categorizethedata
and segregate is effectively. The decision tree approach
utilizes the if-then rules to accomplish its goals of
classification based on the threshold values applied for
execution. The probability scores related to the student
dropout scenario have been subjected to theclassificationto
achieve highly accurate classification of the dropout
predictions. This achieves the student dropout results that
are displayed to the user through the interactive user
interface.
RESULTS AND DISCUSSIONS
The proposed system for student dropout prediction
through the use of Fuzzy Artificial Neural Networkshasbeen
achieved in the Java programming language. The laptop
utilized for the development of this prediction methodology
is equipped with an Intel core i3 processor which is
supplemented by 500 GB of storage and 6 GB of RAM.
For the purpose of achieving the performance metrics of the
proposed methodology the precision and recall parameters
have been assigned. The extraction of the performance of
this approach is necessary to quantify the accuracy of the
student dropout methodology which is being developed
through Fuzzy ANN and decision tree algorithms. The
measurement of the accuracy of the approach will be useful
in determining if these algorithms have been accurately
implemented which will be evident through the scores
achieved by the precision and recall metrics.
Performance Evaluation based on Precision and Recall
The performance metrics through precision and recall have
been assessed through the use of intensive experimentation
that has been performed on the proposed methodology.
These performanceparametersareuseful indetermining the
real accuracy of the evaluationandtheactual performanceof
the methodology. The student dropout detection is highly
useful and can only be viable with a respectable accuracy of
the identification through the methodology prescribed in
this research work.
The in-depth information regarding the performance of the
approach has been achieved throughtheprecisionandrecall
parameters where each of these parameters is useful in
identifying a different aspect of performance of the
methodology. The precision parameter is useful for
determining the relativeperformanceofthestudentdropout
prediction on the other hand recall extracts the absolute
accuracy for the student dropout prediction.
The precision for this methodology has been conceived as
the division of the number of accurate student dropout
predictions by the total number of expectedstudentdropout
predictions. The recall parameter is on the other hand the
absolute accuracy which is derived by the ratio of total
number of student dropout prediction versus the total
number of expected student dropout prediction.
This has been mathematically conveyed through the
equations 4 and 5 given below.
A = The number of correctly predicted student dropout
B= The number of incorrectly predicted student dropout
C = The number of student dropout not predicted
So, precision can be defined as
Precision = (A / (A+ B)) *100
Recall = (A / (A+ C)) *100
The in-depth experimentation done on the prescribed
approach for the measurement of precision and recall
parameters has been listed in the table 1 below.
Table 1: Precision and Recall Measurement Table for
the performance of student dropout prediction](https://image.slidesharecdn.com/273discoveringstudentdropoutpredictionthroughdeeplearning-210721110815/85/Discovering-Student-Dropout-Prediction-through-Deep-Learning-4-320.jpg)
![International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
@ IJTSRD | Unique Paper ID – IJTSRD43700 | Volume – 5 | Issue – 4 | May-June 2021 Page 1553
Figure 2: Comparison of Precision and Recall for the
performance of student dropout prediction
The outcomes achieved in the table are effectively used to
draw a plot graph for precision and recall parameters for
graphical representation in the figure 2 above. As it is
evident from the tabulated values and the outcome of the
line graph, that the precision and recall parametersachieved
for this methodology are well within respectable limits. The
value of precision and recall as 84.06 and 83.54 are
indicative of an accurate implementation of Fuzzy ANN and
Decision Tree approaches that have been utilized for
accurate student dropout prediction in the proposed
methodology.
CONCLUSION
This research article has been effective in the analysis of
predominant researchers on the topic ofdropoutprediction.
The dropouts are highly undesirable occurrencethatusually
is a loss for the student as well as the university involved.
The dropouts can hurt the universities negatively and could
be one of the reasons for the increased dissatisfaction of the
students with the course. These problemscanonlybesolved
through to the effective implementation of a dropout
prediction approach.Thereforethisresearcharticlehasbeen
effective in achieving methodology utilizing machine
learning approaches for the purpose of dropout prediction.
The input dataset is first effectively preprocessed and
labeled before providing it to the K Nearest neighbor for the
clustering. The KNN approach clusters the data using the
certified labels. The generated clusters are provided to the
Fuzzy Artificial Neural Networks as an input which
processes the clusters using the fuzzy crispvaluestoachieve
the predictions. These predictions are then effectively
classified using the if-then rules of the Decision tree
approach. The outcomes indicate the effective
implementation of the student dropout methodology.
The future research directions can be focused on the
realization of the student dropout methodology to be
implemented in a web application for easier access to the
required authorities.
References
[1] T. Hasbun, A. Araya and J. Villalon, "Extracurricular
Activities as Dropout Prediction Factors in Higher
Education Using Decision Trees," 2016 IEEE 16th
International Conference on Advanced Learning
Technologies (ICALT), Austin, TX, 2016, pp. 242-244,
doi: 10.1109/ICALT.2016.66.
[2] D. Sun, Y. Mao, J. Du, P. Xu, Q. Zheng and H. Sun, "Deep
Learning for Dropout Prediction in MOOCs," 2019
Eighth International Conference on Educational
Innovation through Technology (EITT), Biloxi, MS,
USA, 2019, pp. 87-90,doi:10.1109/EITT.2019.00025.
[3] J. Figueroa-Cañas and T. Sancho-Vinuesa, "Early
Prediction of Dropout and Final ExamPerformancein
an Online Statistics Course," in IEEE Revista
Iberoamericana de Tecnologias del Aprendizaje, vol.
15, no. 2, pp. 86-94, May 2020, doi:
10.1109/RITA.2020.2987727.
[4] M. N. Mustafa, L. Chowdhury and M. S. Kamal,
"Students dropout prediction for intelligent system
from tertiary level in developing country," 2012
International Conference on Informatics, Electronics
& Vision (ICIEV), Dhaka, 2012, pp. 113-118, doi:
10.1109/ICIEV.2012.6317441.
[5] L. Wang and H. Wang, "Learning Behavior Analysis
and Dropout Rate Prediction Based on MOOCs Data,"
2019 10th International Conference on Information
Technology in Medicine and Education (ITME),
Qingdao, China, 2019, pp. 419-423, doi:
10.1109/ITME.2019.00100.
[6] N. Shiratori, "Derivation of Student Patterns in a
Preliminary Dropout State and Identification of
Measures for Reducing Student Dropouts," 2018 7th
International Congress on Advanced Applied
Informatics (IIAI-AAI), Yonago, Japan, 2018, pp. 497-
500, doi: 10.1109/IIAI-AAI.2018.00108.
[7] D. Ktoridou and E. Epaminonda, "Measuring the
compatibility between engineering students'
personality types and major of study: A first step
towards preventingengineering educationdropouts,"
2014 IEEE Global Engineering Education Conference
(EDUCON), Istanbul, 2014, pp. 192-195, doi:
10.1109/EDUCON.2014.6826089.
[8] R. Timaran Pereira and J. Caicedo Zambrano,
"Application of Decision Trees for Detection of
Student Dropout Profiles," 2017 16th IEEE
International Conference on Machine Learning and
Applications(ICMLA),Cancun,2017,pp.528-531,doi:
10.1109/ICMLA.2017.0-107.
[9] A. Ortigosa, R. M. Carro, J. Bravo-Agapito, D. Lizcano,J.
J. Alcolea and Ó. Blanco, "From Lab to Production:
Lessons Learnt and Real-Life Challenges of an Early
Student-Dropout Prevention System," in IEEE
Transactions on Learning Technologies, vol. 12,no.2,
pp. 264-277, 1 April-June 2019, doi:
10.1109/TLT.2019.2911608.
[10] K. Murakami et al., "Predicting the Probability of
Student Dropout through EMIR Using Data from
Current and Graduate Students," 2018 7th
International Congress on Advanced Applied
Informatics (IIAI-AAI), Yonago, Japan, 2018, pp. 478-
481, doi: 10.1109/IIAI-AAI.2018.00103.
[11] K. J. de O. Santos, A. G. Menezes, A. B. de Carvalho and
C. A. E. Montesco, "Supervised LearningintheContext
of Educational Data Mining to Avoid University
Students Dropout," 2019 IEEE 19th International
Conference on Advanced Learning Technologies
(ICALT), Macei??, Brazil, 2019, pp. 207-208, doi:
10.1109/ICALT.2019.00068.
[12] R. Lottering, R. Hans and M. Lall, "A model for the
identification of students at risk of dropout at a
university of technology," 2020 International
Conference on Artificial Intelligence, Big Data,
Computing and Data Communication Systems
(icABCD), Durban, South Africa, 2020, pp. 1-8, doi:
10.1109/icABCD49160.2020.9183874.](https://image.slidesharecdn.com/273discoveringstudentdropoutpredictionthroughdeeplearning-210721110815/85/Discovering-Student-Dropout-Prediction-through-Deep-Learning-5-320.jpg)
Discovering Student Dropout Prediction through Deep Learning
- 1. International Journal of Trend in Scientific Research and Development (IJTSRD) Volume 5 Issue 4, May-June 2021 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470 @ IJTSRD | Unique Paper ID – IJTSRD43700 | Volume – 5 | Issue – 4 | May-June 2021 Page 1549 Discovering Student Dropout Prediction through Deep Learning Shashikant Karale, Rajani Pawar, Sharvari Pawar, Poonam Sonkamble Student, Padmabhushan Vasantdada Patil Institute of Technology, Pune, Maharashtra, India ABSTRACT There have been increased incidencesofdropoutthathavebeennoticedinthe universities in the recent years. These increased reports have been instrumental in introducing the graduation rate of the course completionrate for majority of universities all over the globe. Dropouts arehighlyundesirable and are an indication of some underlying inconsistencies that have been plaguing the course since a long time. Therefore, an effective system for the purpose of prediction of the dropout rate is the need of the hour. To reach these goals this research article hasutilizedmachinelearningapproaches.The proposed methodology utilizestheKNearestNeighbor,FuzzyArtificial Neural Network and Decision Tree. This approach has been illustrated in utmost detail in this research article, highlighting the execution of the various important modules of the methodology. The experimentation has been performed to achieve the performance of the approach which has yielded highly accurate results. KEYWORDS: Fuzzy Artificial NeuralNetworks, KNearestNeighbor, DecisionTree, Online Courses How to cite this paper: ShashikantKarale | Rajani Pawar | Sharvari Pawar | Poonam Sonkamble "Discovering StudentDropout Prediction through Deep Learning" Published in International Journal of Trend in Scientific Research and Development(ijtsrd), ISSN: 2456-6470, Volume-5 | Issue-4, June 2021, pp.1549- 1553, URL: www.ijtsrd.com/papers/ijtsrd43700.pdf Copyright © 2021 by author (s) and International Journal ofTrendinScientific Research and Development Journal. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0) (http: //creativecommons.org/licenses/by/4.0) INTRODUCTION Education is an essential and one of the most integral forms of development for a human being. The human being is one of the most intellectually capable organisms on this planet. With the highest brain to body ratio, a lot of activities and other important virtues are the direct result of an increased intellect. As the case with having an intellectual power it needs to be useful and should be honed effectively in a particular way. This honing of the internet is one of themost crucial aspects of growing up. As a child grows up over the years more and more knowledge is accumulated and the child learn new ways and techniques to become self- sustainable and improve his conscience. Education plays an important role in improving the conscience of the individual significantly. Thisimprovement it is done through providing and effective education to the children starting from a very young age.Thisdevelopment is essential for the overall health mentallyandphysicallyofthe child. As the child grows learn more about the world around him and is educated by the parents and at school. School is an important part of a child's life which teaches him a lot of things that would help them later in life and currently. The schools are a heaven of knowledge which imparts valuable education to the children effectively improving their lives over all. After the education provided by the schools these children then effectively are introduced to various different concepts and Fields. And the children utilize these fields and guess their interest to pursue education in those particular interested fields. This in depth education is entrusted form by the Universities and colleges. The University provides education effectively through the use of lecturers and classes. The interested students pursue the relevant course for a period of time stipulated by the University to get proficient in their field. This is essential for the student as well as the educational institution to provide the valuable education to the students. As universities are designed in a capitalistic model they need to make profits to keep providing education to large number of students. But there has been noticed thatthereareincreaseincidences of student dropout from the courses over the past fewyears. These dropouts are highlyproblematicandcanbe effectively detrimental to the university as well as students. The increased numbers of dropouts are basicallyanindicationof some irregularities that are forcing the students to dropout. This process can be ameliorated if the intentions of the students to dropout are effectively visible in time. Therefore there is a need for an effective system for the purpose of predicting the student dropout rate in a particular course. These predictions can be highly valuable and decreasing the dropouts and helping the students complete their course or degree without any hiccups. There have been number of related words that have been analyzed in this research article.These researcheshave been performed to provide resolution for the increasing rates of dropout that are noticed in universities all over the globe. There has been increased interest of the researchers IJTSRD43700
- 2. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470 @ IJTSRD | Unique Paper ID – IJTSRD43700 | Volume – 5 | Issue – 4 | May-June 2021 Page 1550 introducing this problem and finding the root cause of the increased dropout rates. The findings have been effective in understanding that the prediction of a dropout rate if determine for a student it can be effectively useful in addressing the problems and the concerns of the student. This information is valuable in reducing the rate ofdropouts for the students and improving the rate of successful completion for the university. The prediction has also been useful in understanding the inconsistencies in the course material which can be rectified in time before a lot of student’s dropout. Therefore the related words analyzed in this approach have been effective in the designing of our technique for the purpose of dropout predictionthroughthe implementation of machine learning approaches. Literature survey paper provides a thorough analysis of the student dropout rate prediction and realizes an effective machine learning implementation for improving the dropout prediction. In this paper, section 2 is dedicated for the literature review of past work, Section 3 describes the details of the developmental procedure of the model. Section 4 evaluates the results through some experiments and finally section 5 concludes this research article with the traces of the future scope. Literature Review T. Hasbun narrates in Organization for Economic Co- operation and Development countries student dropout is one of the salient problems. [1] Thus for dropout prediction, Educational Data Mining (EDM) analysts are studying institutional interventionsthroughstudents.Intheproposed paper the authors have given importance to extracurricular activities from science degrees engineering and business to predict dropout. The main aim of this paper is to prove that how extracurricular activities can be suitable factors for dropout prediction. The results of the proposed technique have shown better performance. Di Sun states in the last few years there has been a rapid growth in open online courses(MOOCs)butthey'rea serious problem is arising by the MOOC researchersandprovidersis dropout prediction. [2] The paper presents a solution for students getting dropouts just by analyzing how much syllabus can be completed by the students in a course. Thus for predicting the student dropoutrecurrentneural network (RNN) and learning resource representation layer is used thus the accuracy of the proposed papers is higher than the traditional machine learning methodology. J. Canas aims to offer an easy and explainable plan of action to recognize dropout-prone and fail-prone students.[3] For practitioners and researchers studentswhodonotcomplete there higher education or they interrupt their study in between is one of the major concern. The early prediction of the student’s dropout may help teachers to reduce dropout rates and also suggest new learning materials for the students from failing or not completing the course. In the proposed paper the researcher has used tree-based classification models to predict dropout-prone and fail- prone students. M. Mustafa describes for developing countries the national and international loss of student’s dropout prediction is one of the major concerns. The main aim of this paper is to implement an effective dropout prediction system for institutes and colleges.[4] To classify the successful from unsuccessful students they implemented square test methodology on factorssuchasfinancial condition,dropping year, and gender. They have used feature selection, logistic regression, and neural networks to data using data mining techniques Classification and Regression Tree (CART) and CHAm tree. L. Wang explains the emerging wave of artificial intelligence and network information platform technology for the Massive Open Online Courses (MOOC). MOOC is a new classroom model that is greatly hit on the traditional education model.[5] In the proposed paper they implemented a prediction model by using a time-controlled Long Short-TermMemoryneural network (ELSTM).[5]Tasks such as learning evaluation, talent training, curriculum development, teaching mode have greater benefits. Some many shortcomings and problems need to be solved in MOOC. The experimental results of the proposed model are higher which shows the effectiveness of theproposedpaper. N. Shiratori states since the 1990s a serious problem has been are rising in Japan that is dropout ratehasbeena major point of interest by Yomiuri Shimbun. [6] As the semester progresses students lose the interest in course. In preliminary dropout, status distributes the dropout-related state of students by using the logisticregressionmodel.Thus the proposed paper implements model to divide students to examine the issue of student dropout rates. Thus the proposed research students with low motivation in studies so faculty member to know the reasons for this low motivation an early interview. D. Ktoridou narrates that almosthalfofengineeringstudents change their field due to teaching and advising insufficiency, the complexity of the engineering curriculum.[7] There are six personality types enterprising, conventional, artistic, social, and realistic.Theproposedpapershowsthesimilarity between engineering undergraduate students with personality types. This research is done with 113 undergraduate students the main reason for leaving engineering was poor teaching and advising. The paper provides positive effects on both students and Engineering departments R. Pereira presents the student dropout prediction thus by using the data mining techniques disciplinary and institutional data of students from the University of Nariño. [8] By using the classification techniques such as decision trees Socioeconomic and academic student dropout profiles were generated. The proposed model can help to develop policies and strategies for the students to improve the activities for studies. In recent years there have been many types of research made in the field of education by using Data mining techniques. A. Ortigosa states the main aim of the proposed paper is to put a stop to students from leaving the university.[9] The proposed paper collected 11,000students’data fromthefive years by using the C5.0 algorithm which is one of the successful predictive models and SPA i.e Sistema de Prediccioen de Abandono dropout prediction system in Spanish. The SPA model was taken into use in 2017 and is currently in use. Thus by using this model 117,000 risk scores to detect the dropout risk of morethan5700students it to support student dropout risk prevention. K. Murakami performs machine learning techniques to predict dropout rates of current and graduate students by collecting their data thus the main aim of the paper is to
- 3. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470 @ IJTSRD | Unique Paper ID – IJTSRD43700 | Volume – 5 | Issue – 4 | May-June 2021 Page 1551 dropout prevention. [10] The university arranges and collects data of students toprovide enrollmentmanagement. They focus on the student dropout and by using data from both current, graduate, and dropout students. Thus byusing the machine learning algorithm while predicting these dropout rates it is important to investigate unknown items regarding EMIR. K. Santos explains for university student’s dropout prevention by using various educational data mining techniques. [11] The algorithm such as RF and Decision Trees some algorithms are not successful to get high outcomes due to the large data and a parameter tuning step. Thus the proposed paper uses the different supervised learning techniques such as Decision Trees, K-nearest Neighbor, Support Vector Machines, Naive Bayes, Neural Networks, and Random Forests to calculate the prediction models for reducing school dropout. R. Lottering describes in the field of education in South Africa an advanced development has been seen in this development dropout rate is noticed. [12] As in higher education in South Africa student numbers grew by 32.8% from 2006 to 2015 and the dropout rate of 17.1% has been noticed and it can be doubled in the next ten years. Thus proposed systems use data mining techniques and predict dropout goals. The researches of the proposed papers show a positive approach to the proposed model. PROPOSED METHODOLOGY Figure 1: Overview of the Proposed Idea The proposed methodology of the learning outcomes of massive online courses is depicted in figure 1 and it is explained through following steps Step 1: Preprocessing and Labeling – The proposed system uses the dataset of massive online courses forthepurpose of experimentation. The dataset is downloaded from the URL : https://dataverse.harvard.edu/dataset.xhtml?persistentId= doi:10.7910/DVN/29779. Once Testing dataset of the user is received, it is statically stored for the instance learning outcomes. And then proposed model uses the Training set where it reads and stored in a double dimension list, which is subjected to preprocessing thereafter. This double dimension list of dataset belongs to training data is divided into certified and non-certified set based on the Boolean value present in the column of the dataset. Then this dataset is used to label them with the unique integer numerical value using hash set functions. After this all the required rows of the preprocessed list are assigned with the integer numerical value which is then subjected to clustering process in the next step of the model. Step 2: K Nearest Neighbor – The preprocessed list attained in the previous step is utilized as an input in this step of the procedure. This step of the student dropout detection performs the K Nearest Neighbor clustering. This is performed to classify the dataset given as an input into semantic groups to achieve accuratedetectionofthestudent dropout in massive online courses. This procedure is performed through the steps given below. Distance Evaluation – The Euclidean Distance formula given in the equation 1 below is being used for the purpose of evaluation of the distance between the training and testing dataset attributes. The dataset is provided in the form of a list that is given as an input to this step of the procedure. The data stored in the rows of this list is used for the purpose of calculating the distance from the labeled testingdataset,and stored as the row distance at the end of the list. This is performed for all the entries in the list iteratively. ________ (1) Where, ED=Euclidian Distance ATi=Attribute at index i ATj= Attribute at index j Centroid Estimation – The list with the data and the respective row distances calculated in the previous step are used in this step as an input. The list is sorted in the ascending order of the Row distance and subjected to random data point selection. These selected data points are K in number and achieving the centroid. These data points are used to determine the row distance of theselectedindex. The obtained row distance is used further to achieve the boundary of the clusters. The row distances achieved previously are used to form the average row distance of the entire list. These average row distances along with the extracted row distance of the centroid are used to determine the boundaries of the clusters in the cluster formation procedure given below. Cluster Formation – The centroids and the average row distance acquired previously are providedasaninputtothis step for cluster formation. This step then calculates the boundaries of the clusters by addition and subtractionofthe average row distance and the centroidrowdistancetoattain the maximum and minimum values respectively. The data is then subjected to these boundaries to form the clusters which are aggregated into a cluster list. This cluster list sorted into a descending order and the top 3 clusters are provided to the next step as an input. Step 3: Fuzzy Artificial Neural Network – This is one of the most important steps in the proposed methodologywherein a double dimensional list is utilized to effectively store the attributes of the student details as well as the clusters of these values. The fuzzy classification is achieved on this list to effectively segregate the clusters and effectively identify the difference between the maximum and the minimum cluster values. The difference achieved is then effectively divided into 5 equal parts to achieve effective fuzzy classification labels. These labels correspond to the fuzzy crisp values that are segregated as very high medium-low and very low. The student ID and the score sum of different activities are stored in a double list to generate the fuzzy crisp ranges as mentioned in the below algorithm1.
- 4. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470 @ IJTSRD | Unique Paper ID – IJTSRD43700 | Volume – 5 | Issue – 4 | May-June 2021 Page 1552 ________________________________________________ ALGORITHM 1: Fuzzy Crisp Value generation __________________________________________________ //Input : Student Score list SLST //Output: FCLST ( Fuzzy Crisp List) fuuzyCrispValueGeneration(SLST) 1: Start 2: Set min=0.5, max=0.5 3: for i=0 to size of SLST 4: TMPLST = SLST[i] [ TMPLST = Temporary Set] 5: PRSCRE= TMPLST [1] [ Score] 6: IF (PRSCRE < min) 7: min= PRSCRE 8: IF(PRSCRE >max) 9: max= PRSCRE 10: end for 11: RANGE1=0 , RANGE2=0 ,RLIST= ∅ [Rule List] 12: DF=( max-min)/5 [ DF= Diffrence Distance ] 13: for i=0 to 5 14: RANGE1=min 15: RANGE2=RANGE1+DF 16: min=RANGE2 17: TLST[0]= RANGE1 [TLST= Temporary List] 18: TLST[1]= RANGE2 19: FCLST = FCLST + TLST 20: end for 21: return FCLST 22: Stop _________________________________ The neurons for the artificial neural network are then classified according to the given rules above depending on the classified certified set with labels provided such as very high, high, medium, low and very low. These neurons are then subjected to affective evaluationofthehiddenlayerand the resultant probability achievedeffectivelydeterminesthe dropout rate for the desired level according to the user Step 4: Decision Tree – The probability values achieved from the previous step of fuzzy artificial neural networks is taken as an input to this step of the methodology.TheDecisiontree approach is one of the most powerful and highly accurate classification algorithms that completely categorizethedata and segregate is effectively. The decision tree approach utilizes the if-then rules to accomplish its goals of classification based on the threshold values applied for execution. The probability scores related to the student dropout scenario have been subjected to theclassificationto achieve highly accurate classification of the dropout predictions. This achieves the student dropout results that are displayed to the user through the interactive user interface. RESULTS AND DISCUSSIONS The proposed system for student dropout prediction through the use of Fuzzy Artificial Neural Networkshasbeen achieved in the Java programming language. The laptop utilized for the development of this prediction methodology is equipped with an Intel core i3 processor which is supplemented by 500 GB of storage and 6 GB of RAM. For the purpose of achieving the performance metrics of the proposed methodology the precision and recall parameters have been assigned. The extraction of the performance of this approach is necessary to quantify the accuracy of the student dropout methodology which is being developed through Fuzzy ANN and decision tree algorithms. The measurement of the accuracy of the approach will be useful in determining if these algorithms have been accurately implemented which will be evident through the scores achieved by the precision and recall metrics. Performance Evaluation based on Precision and Recall The performance metrics through precision and recall have been assessed through the use of intensive experimentation that has been performed on the proposed methodology. These performanceparametersareuseful indetermining the real accuracy of the evaluationandtheactual performanceof the methodology. The student dropout detection is highly useful and can only be viable with a respectable accuracy of the identification through the methodology prescribed in this research work. The in-depth information regarding the performance of the approach has been achieved throughtheprecisionandrecall parameters where each of these parameters is useful in identifying a different aspect of performance of the methodology. The precision parameter is useful for determining the relativeperformanceofthestudentdropout prediction on the other hand recall extracts the absolute accuracy for the student dropout prediction. The precision for this methodology has been conceived as the division of the number of accurate student dropout predictions by the total number of expectedstudentdropout predictions. The recall parameter is on the other hand the absolute accuracy which is derived by the ratio of total number of student dropout prediction versus the total number of expected student dropout prediction. This has been mathematically conveyed through the equations 4 and 5 given below. A = The number of correctly predicted student dropout B= The number of incorrectly predicted student dropout C = The number of student dropout not predicted So, precision can be defined as Precision = (A / (A+ B)) *100 Recall = (A / (A+ C)) *100 The in-depth experimentation done on the prescribed approach for the measurement of precision and recall parameters has been listed in the table 1 below. Table 1: Precision and Recall Measurement Table for the performance of student dropout prediction
- 5. International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470 @ IJTSRD | Unique Paper ID – IJTSRD43700 | Volume – 5 | Issue – 4 | May-June 2021 Page 1553 Figure 2: Comparison of Precision and Recall for the performance of student dropout prediction The outcomes achieved in the table are effectively used to draw a plot graph for precision and recall parameters for graphical representation in the figure 2 above. As it is evident from the tabulated values and the outcome of the line graph, that the precision and recall parametersachieved for this methodology are well within respectable limits. The value of precision and recall as 84.06 and 83.54 are indicative of an accurate implementation of Fuzzy ANN and Decision Tree approaches that have been utilized for accurate student dropout prediction in the proposed methodology. CONCLUSION This research article has been effective in the analysis of predominant researchers on the topic ofdropoutprediction. The dropouts are highly undesirable occurrencethatusually is a loss for the student as well as the university involved. The dropouts can hurt the universities negatively and could be one of the reasons for the increased dissatisfaction of the students with the course. These problemscanonlybesolved through to the effective implementation of a dropout prediction approach.Thereforethisresearcharticlehasbeen effective in achieving methodology utilizing machine learning approaches for the purpose of dropout prediction. The input dataset is first effectively preprocessed and labeled before providing it to the K Nearest neighbor for the clustering. The KNN approach clusters the data using the certified labels. The generated clusters are provided to the Fuzzy Artificial Neural Networks as an input which processes the clusters using the fuzzy crispvaluestoachieve the predictions. These predictions are then effectively classified using the if-then rules of the Decision tree approach. The outcomes indicate the effective implementation of the student dropout methodology. The future research directions can be focused on the realization of the student dropout methodology to be implemented in a web application for easier access to the required authorities. References [1] T. Hasbun, A. Araya and J. Villalon, "Extracurricular Activities as Dropout Prediction Factors in Higher Education Using Decision Trees," 2016 IEEE 16th International Conference on Advanced Learning Technologies (ICALT), Austin, TX, 2016, pp. 242-244, doi: 10.1109/ICALT.2016.66. [2] D. Sun, Y. Mao, J. Du, P. Xu, Q. Zheng and H. Sun, "Deep Learning for Dropout Prediction in MOOCs," 2019 Eighth International Conference on Educational Innovation through Technology (EITT), Biloxi, MS, USA, 2019, pp. 87-90,doi:10.1109/EITT.2019.00025. [3] J. Figueroa-Cañas and T. Sancho-Vinuesa, "Early Prediction of Dropout and Final ExamPerformancein an Online Statistics Course," in IEEE Revista Iberoamericana de Tecnologias del Aprendizaje, vol. 15, no. 2, pp. 86-94, May 2020, doi: 10.1109/RITA.2020.2987727. [4] M. N. Mustafa, L. Chowdhury and M. S. Kamal, "Students dropout prediction for intelligent system from tertiary level in developing country," 2012 International Conference on Informatics, Electronics & Vision (ICIEV), Dhaka, 2012, pp. 113-118, doi: 10.1109/ICIEV.2012.6317441. [5] L. Wang and H. Wang, "Learning Behavior Analysis and Dropout Rate Prediction Based on MOOCs Data," 2019 10th International Conference on Information Technology in Medicine and Education (ITME), Qingdao, China, 2019, pp. 419-423, doi: 10.1109/ITME.2019.00100. [6] N. Shiratori, "Derivation of Student Patterns in a Preliminary Dropout State and Identification of Measures for Reducing Student Dropouts," 2018 7th International Congress on Advanced Applied Informatics (IIAI-AAI), Yonago, Japan, 2018, pp. 497- 500, doi: 10.1109/IIAI-AAI.2018.00108. [7] D. Ktoridou and E. Epaminonda, "Measuring the compatibility between engineering students' personality types and major of study: A first step towards preventingengineering educationdropouts," 2014 IEEE Global Engineering Education Conference (EDUCON), Istanbul, 2014, pp. 192-195, doi: 10.1109/EDUCON.2014.6826089. [8] R. Timaran Pereira and J. Caicedo Zambrano, "Application of Decision Trees for Detection of Student Dropout Profiles," 2017 16th IEEE International Conference on Machine Learning and Applications(ICMLA),Cancun,2017,pp.528-531,doi: 10.1109/ICMLA.2017.0-107. [9] A. Ortigosa, R. M. Carro, J. Bravo-Agapito, D. Lizcano,J. J. Alcolea and Ó. Blanco, "From Lab to Production: Lessons Learnt and Real-Life Challenges of an Early Student-Dropout Prevention System," in IEEE Transactions on Learning Technologies, vol. 12,no.2, pp. 264-277, 1 April-June 2019, doi: 10.1109/TLT.2019.2911608. [10] K. Murakami et al., "Predicting the Probability of Student Dropout through EMIR Using Data from Current and Graduate Students," 2018 7th International Congress on Advanced Applied Informatics (IIAI-AAI), Yonago, Japan, 2018, pp. 478- 481, doi: 10.1109/IIAI-AAI.2018.00103. [11] K. J. de O. Santos, A. G. Menezes, A. B. de Carvalho and C. A. E. Montesco, "Supervised LearningintheContext of Educational Data Mining to Avoid University Students Dropout," 2019 IEEE 19th International Conference on Advanced Learning Technologies (ICALT), Macei??, Brazil, 2019, pp. 207-208, doi: 10.1109/ICALT.2019.00068. [12] R. Lottering, R. Hans and M. Lall, "A model for the identification of students at risk of dropout at a university of technology," 2020 International Conference on Artificial Intelligence, Big Data, Computing and Data Communication Systems (icABCD), Durban, South Africa, 2020, pp. 1-8, doi: 10.1109/icABCD49160.2020.9183874.