1

Int. J. Mech. Eng. Res. & Tech 2023

Int. J. Mech. Eng. Res. & Tech 20221

ISSN 2454 – 535X www.ijmert.com

Vol. 11, Issue.3, July 2023

Using Machine Learning To Predict Student Performance In Online

Learning Environments
By

Raunak Bedi
Abstract

Prediction of academic performance of students is one of the major topics for universities and
schools as it can be helpful to design the right mechanisms to avoid dropouts and improve
academic results, among others. A lot of processes have been automated in usual activities of
students to benefit them and manage big data gathered from software products for tech-based
learning. Hence, processing and analyzing the same data properly can give a lot of vital insights
to their knowledge and relation between students and their homework. This information can feed
promising methods and algorithms for prediction of student performance.
This study is conducted to analyze various machine learning models used for predicting student’s
performance. This study presents an in-depth review of studies examining data of online learning
environments to predict students’ outcomes with machine learning techniques. This study will help
identify the online course features used for prediction of learners’ outcome, determine the outputs
of prediction, strategies, and methodologies of feature extraction for prediction of performance,
evaluation metrics, and challenges and limitations for analyzing the outcomes.

Keywords: academic performance, student performance, online learning, machine learning

algorithms, machine learning models, online learning environments

Introduction

The way people used to learn has been revolutionized by online learning and education has never
been so convenient and affordable to billions of people worldwide. Irrespective of rising interest
and benefits of distance and online learning, universities and schools are highly concerned about
students’ retention and academic performance, along with low degree/certification completion
rates and high dropout rates.
Shalom Presidency School
st
A-199 1 floor sector 55 Gurgaon, Haryana,122011, India
bedi.raunak@gmail.com
Phone No=+91 9354980255

2
 algorithm can determine whether an image
Dropping out or failing an online program has object(Watt et al, 2020).
or course is usually an important parameter
to assess course/program quality and
allocate resources by institutional Additionally, machine learning can
authorities. understand the products a customer might
like. After analyzing the previous products,
the system suggests new product that might
Low certification and dropout rates are also be interesting to the customer (Witten et al,
a major risk factor to profitability, funding, 2016). All such examples have similar
and reputation of an institution (Arce et al, principle. The data is processed by the
2015). These outcomes have vast impact system and this data can be identified, and
on well-being, self-esteem, odds of then this knowledge is used to make future
graduating, and employment of students decisions. The rise in data has been
(Arce et al., 2015; Larusson & White, 2014). effective to make such applications more
Hence, it is important to find more efficient effective. Machine learning is categorized
methods to forecast performance of into supervised and unsupervised learning
students as early as possible for students, as per the type of input. Input data belongs
educators, and institutions to take necessary to a common class structure in supervised
measures for improving online learning learning (Mitchell, 2007; Kumar et al,
experiences of students and building 2012). This input data is called training
intervention strategies to meet the needs of data. The algorithm is basically aimed to
students. With rising interest of online create a prediction model for predicting a
learning and big data produced by students property with other properties. After
by interacting with online learning creating the model, it processes data with
environments, several machine learning similar class structure to input data. There is
methods are proposed by researchers to no known class structure in input data and
predict students’ performance and improve algorithm is aimed to reveal the data
their learning outcomes. structure in unsupervised learning
1.1 Background (Mitchell, 1997; Sugiyama, 2015).

Machine learning is applied in different Literature Reviews

areas. For example, machine learning is
used in search engines to create relations
between webpages and search terms. Student retention is said to be a major
Search engines scan the website’s content concern in education. Though intervention
and define the most important terms and strategies can improve retention rates, prior
words to define a specific webpage and use knowledge of students’ performance for
the same information to return the most those programs (Yadav et al., 2012). This
relevant information to a given search term is where it is important to predict student
(Witten et al., 2016). Machine learning is performance. Using machine learning for
also used in image recognition to identify performance prediction or dropout is a
specific objects like faces (Alpaydin, common pattern in academic studies. In
2020). Initially, machine learning model online learning, dropout prediction is a very
analyzes images with a specific object. If common concern in those studies because
enough images are given for processing, the of both easy availability of data and high

3
dropout rates (Kalles and Pierrakeas, machine learning models, they also used
2006). Areas out of online learning are ensembled techniques and compared the
major contexts where performance or results. They found best result with
dropout predictions are widely used for decision trees. Behavioral features were the
research purposes. Purpose of these studies other area focused by the researchers. They
varies. Finding the best prediction created a model by taking these features or
approach is important in some studies. without them. The prediction results were
Some studies are aimed to determine the improved by including these behavioral
viability of machine learning to predict features.
student performance ordropout.
Cortez & Silva (2008) conducted a study
A study was conducted at the “Eindhoven on performance prediction at the
University of Technology” to determine “University of Minho, Portugal”. The
the effectiveness of machine learning to dataset had information on whether exam
predict students’ dropout rate (Dekker et has been passed by the students in
al., 2009). Building various prediction Portuguese language and Mathematics.
models with various machine learning They used ML models like random forest,
approaches like Logit, BayesNet and decision trees, support vector machines,
CART are the basic methodology here. and neural networks and compared them
Then, they compared prediction outcomes for accuracy. They also compared a dataset
of various models in terms of effectiveness. with previous exam results and the one
J48 classifier has successfully built the which didn’t have past grades.
most efficient model. A group of Performance was improved by adding
researchers from three Indian universities previous results.
have conducted a similar study. They
analyzed the dataset of university students
using various algorithms. They compared 2.1 Research Gap
the recall value and precision later. They E-learning has become a common form of
got most accurate results with the ADT education and a vital part of development
decision tree algorithm (Yadav et al.,2012). of online education (Giannakos & Vlamos,
However, prediction of performance of 2013). E-learning has become a common
students rather than dropouts is more phenomenon because of the impact of
relevant with this study. There are also COVID-19 across the world because of its
some studies which have predicted spatial flexibility, high temporal, rich
students’ performance. In “Hellenic Open education materials, and low learning
University”, a study was done to analyze curve. However, teachers cannot perceive
the use of ML in distance education by learning status of the students easily in this
Kalles & Pierrakeas (2006). They used mode (Qu et al., 2019). Hence, it raises
decision trees and genetic algorithms to concerns over the quality of online
come up with a predictive model and learning. Hence, this study is based on
compared the results for accuracy. The education, i.e., by predicting student
“Genetically Evolved Decision Trees performance. It also evaluates
(GATREE)” model has provided most effectiveness of various machine learning
accurate results. approaches. Recall, precision, F-

Amrieh et al (2016) conducted a study on

predicting performance at the “University
of Jordan”. They used a dataset of students
from various nations. Along with separate
4
measure, etc. are some of the most to various machine learning techniques to
common indicators to determine the predict student performance (out of which
effectiveness of machine learning models 90% of the studies were published over the
(Powers, 2020). past two decades). The studies selected for
this paper are published in various journals,
conferences, and book chapters. The
2.2 Research Question corresponding literature has been extracted
from various online databases like IEEE,
 What are the effective machine learning Springer, Science Direct, ACM Digital
models to predict students’ performance Library, iJET, Sage Journals, Wiley Online
in e-learning? Library, Google Scholar, etc.

2.3 Research Objectives

 To predict student’s performance in
online learning with machine learning
 To evaluate various machine learning
models for performance prediction

Research Methodology

The study of performance prediction

provides a foundation to teachers for
adjusting their approaches of teaching for
students who may have trouble by
predicting their performance on upcoming
exams. It can help reduce the risk of failure
during the course and improve the quality
of online education. With a lot of empirical
studies investigating the relation between
learning performance and e-learning
behavior, the e-learning behavior plays a
vital role on learning outcome. Research
community has focused a lot on prediction
of learning performance on the basis of
data on learning process in recent years.
Teachers can modify their teaching
strategies over time with the help of
collection, analysis, and measurement of
data on learning process to predict learning
performance and start using early warning
and supervision during the learning
processes.

This study is qualitative in nature as around

70 recent studies have been selected related
5
Exclusion criteria for this study includes certificate prediction are the same. A
papers without proper contribution or “certificate prediction model” has been
having lack of quality or clarity. Research developed by Kórösi et al. (2018) on the
papers without impact factor or not peer- basis of features associated with mouse
reviewed have been excluded. The papers behavior, learning behavior, text inputs,
published in conferences not supported or and video-watching attributes. Various
published by Springer, IEEE, ACM or other classification models are used to implement
reputed editorials or organizations were the “gain-ratio feature selection” method
excluded as well. and bagging and random forest are used to
obtain the best performance.
Analysis of Study
In the same way, “Al-Shabandar et al.
(2017)” have proposed hill climbing and
There are various reasons researchers have random forest model to choose statistical
learned students’ learning behavior and features automatically from behavioral and
characteristics, such as understanding demographical attributes of students. The
learning styles of students, building “HarvardX (2014)” dataset was used to
students’ profiles to provide personalized under-sample majority class to solve the
learning, and boosting educational problem related to class imbalance. A
outcomes. They have investigated different “deep neural network” model has been
outcomes which can be grouped into proposed by Imran et al (2019) for the
completion and retention predictions as prediction of certificate acquisition and
well as performance predictions as student dropout on the basis of learner
retention and success rate are important behavioral data.
variables to assess and measure constantly
in institutions. The performance prediction
is done to predict the efficiency of students An approach has been proposed by Liang
on completing the given course like grade et al. (2017) to provide personalized profile
value, grade category, failure/success, to boost e- learning to manage student
certificate, and risk prediction. A lot of behavior. They explored various
researchers have studied “student dropout classification models for prediction of
prediction” in this field. whether a students will get a certificate on
the basis of “Jaccard coefficient similarity”
of learning behavior and student profiles.
4.1. Predicting Student’s They tested the model over 5 to 7 weeks
Performance in Online and obtained best results at 7 weeks with the
SVM model. An approach has been
Learning There are three proposed by Ruipérez-Valiente et al. (2017)
performance measures examined in which uses “statistical learning behavior”
recent studies for prediction of student and progress features with various
performance – certificate acquisition, grade approaches of machine learning to predict
prediction, and student at-risk prediction. whether a certificate will be given to a
student or not.
4.1.1. Certificate Acquisition
Certificate acquisition of student is one of 4.1.2. Grades
the best measures to predict performance of
students in professional courses. In this It is yet another perspective on success for
category, all the studies have predicted any student. A lot of publications have
achievement of certificate of completion. proposed prediction models for grades of
Hence, achieving course completion and students in different assessments like final-
6
exam grades, course grades, assignments, statistical features like % of total accurate
or quizzes in the course. Researchers have answers and total time invested on video
tested regression and classification models resources from data and combined the same
to predict grades. Binary classification has with weights which were previously
been used in various studies for prediction learned from former courses for prediction
of success and failure of students (Huang et of “students-at- risk” and protecting them
al, 2020; Liu et al., 2017; Wan et al., 2019; from failing in existing course. A model has
Lemay & Doleck, 2020; Kokoç et al., been proposed by “El Aouifi et al (2021)”
2021). Various machine learning models for prediction of final grade on the basis of
have been examined in mainstream studies interactions of students with pedagogical
with manually selected features of sequences of behaviors related to
statistical learning behavior to predict educational video like jump forward, play,
grades of students. Xiao et al. (2018) used pause, jump back, and end. They fathered
“classification and regression tree sequences as features as per educational
(CART)” model on learning behavior data sequences. They used MLP and K-NN to
and statistical demographic for the achieve best results.
classification of final grades of students into
multiple classes. Additionally, a machine
learning model has been proposed by 4.2. Machine Learning
Villagrá- Arnedo et al. (2016) which Models for
used normalized behavior data for Performance
classification of students in 5 categories
Prediction
of grades. Here, “support vector machine
(SVM)” model performed better than other A huge range of deep learning and machine
baseline models. learning techniques have been used by
researchers for prediction of outcomes of e-
learning from the data of students’ online
4.1.3. Predicting Students-at- interaction. In this field, the machine
risk learning models are categorized into linear
models like “logistic regression”,
Along with predicting performance, it is
probabilistic (Naïve Bayes), tree-based
also important to identify students who
models like “decision tree (DT)”, ensemble
might be failed in a course. Kondo et al
model like AdaBoost, linear model like
(2017) identified at-risk and off-task
“logistic regression (LR)”, instance-based
students using “attendance attributes” and
model like “K-nearest neighbor (kNN)”,
“learning interaction” of “SPOCs dataset”.
and rule-based models like “fuzzy logic”
Cano & Leonard (2019) proposed a
approaches. Additionally, some studies
“genetic programming (GP)” based “early
have used heuristic or optimization
warning system to address socioeconomic
methods like K-star model for prediction of
disadvantages of students. It extracts
outcomes.
classification rules automatically on the
basis of learning interaction, student Recent studies have investigated deep
demographic, registration, academic learning-based models like “long short-
background, and socioeconomic and family term memory (LSTM)” and “convolutional
data. neural networks (CNNs)” to predict e-
learning outcomes. Table 1 lists all the
models based on deep learning and
A transfer-based model has been proposed machine learning used in earlier studies.
by Wan et al. (2019) to identify “at-risk Figure 1 illustrates the categories of
students” on a class test which is held every predictive models in relevant literature.
week. They extracted behavioral and Deep learning models are used most
7
commonly in recent studies. It is also worth developing the “feed-forward neural
considering that “deep learning” is the term network” with over 3 hidden layers are
which refers to models implementing neural considered as deep learning.
network with over 3 layers. Hence, studies

Category Machine Learning Models

Linear Models “Logistic regression (LR), support vector machine (SVM), linear
discriminant analysis (LDA), generalized linear model (GLM), lasso
linear regression (LLG), boosted logistic regression”
“Probabilistic” “Naive Bayes, Bayes network, Bayesian generalized linear (BGL),
Bayesian belief networks”
“Tree-based models” “Decision tree (DT), random forest (RF), Bayesian additive regression
trees (BART)”
Instance-based “k-nearest neighbors (kNN)”
Rule-Based “Rule-based classifier (JRip), fuzzy set rules”
Neural network “Multilayer perceptron(MLP) or artificial neural network (ANN)”
Deep neural network “Recurrent neural network (RNN), gated recurrent unit (GRU), long
short-term memory (LSTM), convolutional neural network (CNN),
squeeze-and-excitation networks (SE-net)”
Table 1 – Machine learning models and their categories used in previous studies

Figure 1 – Machine learning models used in previous studies

Source – Alhothali et al. (2022)

Recommendation systems gather data on user choices for a range of elements like

8
websites, books, applications, travel prediction model for academic
destinations, e-learning, etc. In context of performance (Iyanda et al., 2018). Finally,
performance of students, it is possible to ANN’s potential for prediction of learning
acquire explicit information or implicit outcomes is compared to “multivariate LR
information by gathering scores and model” in medical education
monitoring their behavior like materials (Dharmasaroja & Kingkaew, 2016).
downloaded and visiting study materials
(Bobadilla et al, 2013). Recommender Results
systems consider various data sources for
predictions. They manage factors like
novelty, precision, stability, and dispersion. Application of techniques like
4.2.1. Collaborative Filtering collaborative filtering, machine learning,
artificial neural network and recommender
It plays a vital role in prediction, even systems can consider different types of
though it is used with other filtering information for students’ behavioral
methods like knowledge-based, content- prediction, for example, tasks’ grades and
based, or social approaches (Bobadilla et demographic characteristics. A study by
al, 2013). Just like decisions are based as the “Hellenic Open University” is a good
per previous knowledge and experiences, point to start, where researchers applied
collaborative filtering is used to perform various supervised and
prediction. Some studies have predicted
various issues regarding performance of
students with collaborative filtering.
Hence, Bydžovská (2015) found
similarities between students where their
knowledge was represented as a range of
grades from earlier courses.

4.2.2. “Artificial Neural Networks (ANN)”

ANN includes a range of entities which are
highly connected internally like
“Processing Elements”. The function and
structure of the network is inspired by
human brain. Each element of processing
is designed to work like its neuron, which
accepts weighted inputs and responds by
giving output (Adewale et al., 2018). ANN
is applied in various prediction models,
usually by taking in evaluation outcomes of
students. The scores of tests were predicted
with a feedforward ANN by taking in
partial scores (Gedeon & Turner, 1993).
The academic performance is predicted by
the ANN which uses “Cumulative Grade
Point Average” in 8th semester (Arsad &
Buniyamin, 2013). Researchers compared
two ANN models, i.e., “Generalized
Regression Neural Network” and
“Multilayer Perceptron” to identify the best
9
machine learning models to a specific techniques like social, content-based and
dataset. In this study, it is observed that knowledge based (Bobadilla et al., 2013).
“Naïve Bayes” model was the best Just like decisions made by humans are
algorithm to predict both probability and based on their knowledge and past
performance of dropout of students experience, collaborative filtering acts just
(Kotsiantis et al., 2004). like the same for predictions. Some studies
predicted various issues regarding the
performance of students with collaborative
Machine learning consists of techniques filtering. Majority of studies discussed in
which enables computers to learn without this article use large data metrices for
human intervention (Navamani & performance prediction of students. Hence,
Kannammal, 2015). Machine learning has collaborative filtering was not that accurate
helped in different applications like stock in prediction when used in small sample
market analysis, medical diagnostics, sizes for the same purpose (Pero &
classification of DNA sequence, robotics, Horváth, 2015).
games, and predictive analysis. Rastrollo-
Guerrero et al. (2020) focused on
predictive analysis where they implemented
complex models for prediction. These
approaches can be helpful to promote
decision-making with relevant data. Conclusion
Supervised learning promotes models to
reason from cases which are supplied
externally to generate hypothesis which
On the basis of data collected here,
can make predictions on future events
supervised learning is the most common
(Kotsiantis, 2007).
technique for student behavioral prediction
as it provides reliable and accurate results.
All in all, supervised learning is mainly The SVM model was used most widely by
aimed to come up with a clear distribution the authors and has given most accurate
model of class labels in predictor features. results. Along with SVM, decision tree,
Rule indication is an excellent supervised naïve bayes, and random forests are widely
learning approach to make predictions used for good results. Collaborative
which can reach 94% accuracy level when filtering algorithms and recommender
it comes to predict new student dropouts in systems are widely used in this arena. It is
nursing (Moseley & Mead, 2008). It is worth noting that recommending activities
important to maintain caution with and resources are more successful than
classification techniques and look for any prediction of student behavior. Neural
unbalanced datasets as they can mislead networks are not used widely but they
when it comes to prediction. This way, achieved great accuracy when it comes to
Nandeshwar et al. (2011) proposed various predict performance of students. This study
improvements for prediction of student can help researchers and give them a lot of
dropout like exploring a huge range of possibilities to apply machine learning for
methods, attributes, and evaluating theory prediction of performance in e-learning and
effectiveness and studying factors among other platforms.
non-dropout and dropout students.

Collaborative filtering approaches play a

vital role in recommendation, even though References
they usually go hand in hand with other Adewale, A. M., Bamidele, A. O., &
10
Lateef, U. O. (2018). Predictive modelling data mining to predict secondary school
and analysis of academic performance of student performance. Proceedings of 5th
secondary school students: Artificial Annual Future Business Technology
Neural Network approach. International Conference, Porto, 5-12.
Journal of Science and Technology Dekker, G. W., Pechenizkiy, M., &
Education Research, 9(1), 1-8. Vleeshouwers, J. M. (2009). Predicting
Alpaydin, E. (2020). Introduction to Students Drop Out: A Case Study.
machine learning. MIT press. International Working Group on
Al-Shabandar, R., Hussain, A., Laws, A., Educational Data Mining.
Keight, R., Lunn, J., & Radi, N. (2017, Dharmasaroja, P., & Kingkaew, N. (2016,
May). Machine learning approaches to August). Application of artificial neural
predict learning outcomes in Massive open networks for prediction of learning
online courses. In 2017 International joint performances. In 2016 12th International
conference on neural networks (IJCNN) Conference on Natural Computation,
(pp. 713-720). IEEE. Fuzzy Systems and Knowledge Discovery
Amrieh, E. A., Hamtini, T., & Aljarah, I. (ICNC-FSKD) (pp. 745-751). IEEE.
(2016). Mining educational data to predict El Aouifi, H., El Hajji, M., Es-Saady, Y.,
student’s academic performance using & Douzi, H. (2021). Predicting learner’s
ensemble methods. International journal of performance through video sequences
database theory and application, 9(8), 119- viewing behavior analysis using
136. educational data-mining. Education and
Arce, M. E., Crespo, B., & Míguez- Information Technologies, 26(5), 5799-
Álvarez, C. (2015). Higher Education 5814.
Drop-out in Spain-- Particular Case of Gedeon, T. D., & Turner, H. S. (1993,
Universities in Galicia. International October). Explaining student grades
Education Studies, 8(5), 247-264. predicted by a neural network. In
Arsad, P. M., & Buniyamin, N. (2013, Proceedings of 1993 International
November). A neural network students' Conference on Neural Networks (IJCNN-
performance prediction model (NNSPPM). 93- Nagoya, Japan) (Vol. 1, pp. 609-612).
In 2013 IEEE International Conference on IEEE.
Smart Instrumentation, Measurement and Giannakos, M. N., & Vlamos, P. (2013).
Applications (ICSIMA) (pp. 1-5). IEEE. Educational webcasts' acceptance:
Bobadilla, J., Ortega, F., Hernando, A., & Empirical examination and the role of
Gutiérrez, A. (2013). Recommender experience. British Journal of Educational
systems survey. Knowledge-based systems, Technology, 44(1), 125-143.
46, 109-132. HarvardX. (2014). HarvardX Person-
Bydžovská, H. (2015). Student Course Academic Year 2013 De-Identified
performance prediction using Dataset, Version3.0.
collaborative filtering methods. In Huang, A. Y., Lu, O. H., Huang, J. C., Yin,
Artificial Intelligence in Education: 17th C. J., & Yang, S. J. (2020). Predicting
International Conference, AIED 2015, students’ academic performance by using
Madrid, Spain, June 22-26, 2015. educational big data and learning analytics:
Proceedings 17 (pp. 550-553). Springer evaluation of classification methods and
International Publishing. learning logs. Interactive Learning
Cano, A., & Leonard, J. D. (2019). Environments, 28(2), 206-230.
Interpretable multiview early warning Imran, A. S., Dalipi, F., & Kastrati, Z.
system adapted to underrepresented student (2019, April). Predicting student dropout in
populations. IEEE Transactions on a MOOC: An evaluation of a deep neural
Learning Technologies, 12(2), 198-211. network model. In Proceedings of the 2019
Cortez, P., & Silva, A. M. G. (2008). Using 5th International Conference on
11
Computing and Artificial Intelligence (pp. Learning analytics: From research to
190-195). practice (Vol. 13). Springer.
Iyanda, A. R., Ninan, O. D., Ajayi, A. O., & Lemay, D. J., & Doleck, T. (2020). Grade
Anyabolu, O. G. (2018). Predicting Student prediction of weekly assignments in
Academic Performance in Computer MOOCS: mining video-viewing behavior.
Science Courses: A Comparison of Education and Information Technologies,
Neural Network Models. International 25, 1333-1342.
Journal of Modern Education & Computer Liang, K., Zhang, Y., He, Y., Zhou, Y.,
Science, 10(6). Tan, W., & Li, X. (2017). Online behavior
Kalles, D., & Pierrakeas, C. (2006). analysis-based student profile for intelligent
Analyzing student performance in distance E-learning. Journal of Electrical and
learning with genetic algorithms and Computer Engineering, 2017. Liu, W., Wu,
decision trees. Applied Artificial J., Gao, X., & Feng, K. (2017, December).
Intelligence, 20(8), 655-674. An early warning model of student
Kokoç, M., Akçapınar, G., & Hasnine, M. achievement based on decision trees
N. (2021). Unfolding students’ online algorithm. In 2017 IEEE 6th International
assignment submission behavioral patterns Conference on Teaching, Assessment, and
using temporal learning analytics. Learning for Engineering (TALE) (pp. 517-
Educational Technology & Society, 24(1), 222). IEEE.
223-235. Mitchell, T. M. (2007). Machine learning
Kondo, N., Okubo, M., & Hatanaka, T. (Vol. 1). New York: McGraw-hill.
(2017, July). Early detection of at-risk
Moseley, L. G., & Mead, D. M. (2008).
students using machine learning based on
Predicting who will drop out of nursing
LMS log data. In 2017 6th IIAI
courses: a machine learning exercise. Nurse
international congress on advanced
education today, 28(4), 469-475.
applied informatics (IIAI-AAI) (pp. 198-
Nandeshwar, A., Menzies, T., & Nelson,
201). IEEE.
A. (2011). Learning patterns of university
Kórösi, G., Esztelecki, P., Farkas, R., &
student retention. Expert Systems with
Tóth, K. (2018). Clickstream-based
Applications, 38(12), 14984-14996.
outcome prediction in short video MOOCs.
Navamani, J. M. A., & Kannammal, A.
In 2018 International conference on
(2015). Predicting performance of schools
computer, information and
by applying data mining techniques on
telecommunication systems (CITS) (pp. 1-
public examination results. Research
5). IEEE.
Journal of Applied Sciences, Engineering
Kotsiantis, S. (2007). A review of
and Technology, 9(4), 262-271.
classification techniques, sb kotsiantis.
Pero, Š., & Horváth, T. (2015).
Informatica, 31, 249- 268.
Comparison of collaborative-filtering
Kotsiantis, S., Pierrakeas, C., & Pintelas, P. techniques for small-scale student
(2004). PREDICTING performance prediction task. In
STUDENTS'PERFORMANCE IN Innovations and Advances in Computing,
DISTANCE LEARNING USING Informatics, Systems Sciences, Networking
MACHINE LEARNING and Engineering (pp. 111-116). Springer
TECHNIQUES. Applied Artificial International Publishing. Powers, D. M.
Intelligence, 18(5), 411-426. (2020). Evaluation: from precision, recall
Kumar, S., Mohri, M., & Talwalkar, A. and F-measure to ROC, informedness,
(2012). Sampling methods for the Nyström markedness and correlation. arXiv preprint
method. The Journal of Machine Learning arXiv:2010.16061.
Research, 13(1), 981-1006. Qu, S., Li, K., Wu, B., Zhang, X., & Zhu, K.
Larusson, J. A., & White, B. (Eds.). (2014). (2019). Predicting student performance and
deficiency in mastering knowledge points in
12
MOOCs using multi-task learning. 2018 International Conference on Sensor
Entropy, 21(12), 1216. Networks and Signal Processing (SNSP)
-Guerrero, J. L., Gómez-Pulido, J. A., & (pp. 250-254). IEEE.
Durán-Domínguez, A. (2020). Analyzing Yadav, S. K., Bharadwaj, B., & Pal, S.
and predicting students’ performance by (2012). Mining Education data to predict
means of machine learning: A review. student's retention: a comparative study.
Applied sciences, 10(3), 1042. arXiv preprint arXiv:1203.2987.
Ruipérez-Valiente, J. A., Cobos, R.,
Muñoz-Merino, P. J., Andujar, Á., &
Delgado Kloos, C. (2017). Early
prediction and variable importance of
certificate accomplishment in a MOOC.
In Digital Education: Out to the World and
Back to the Campus: 5th European
MOOCs Stakeholders Summit, EMOOCs
2017, Madrid, Spain, May 22-26, 2017,
Proceedings 5 (pp. 263- 272). Springer
International Publishing.
Sugiyama, M. (2015). Introduction to
statistical machine learning. Morgan
Kaufmann.
Villagrá-Arnedo, C. J., Gallego-Durán, F. J.,
Compañ, P., Llorens Largo, F., & Molina-
Carmona,
R. (2016). Predicting academic performance
from behavioural and learning data.
Wan, H., Liu, K., Yu, Q., & Gao, X.
(2019). Pedagogical intervention practices:
Improving learning engagement based on
early prediction. IEEE Transactions on
Learning Technologies, 12(2), 278-289.
Wan, H., Liu, K., Yu, Q., & Gao, X.
(2019). Pedagogical intervention practices:
Improving learning engagement based on
early prediction. IEEE Transactions on
Learning Technologies, 12(2), 278-289.
Watt, J., Borhani, R., & Katsaggelos, A. K.
(2020). Machine learning refined:
Foundations, algorithms, and applications.
Cambridge University Press.
Witten, I. H., Frank, E., Hall, M. A., Pal, C.
J., & DATA, M. (2016, June). Practical
machine learning tools and techniques. In
Data Mining (Vol. 2, No. 4).
Xiao, B., Liang, M., & Ma, J. (2018,
October). The application of CART
algorithm in analyzing relationship of
MOOC learning behavior and grades. In

13