Technology, Knowledge and Learning (2019) 24:567–598

https://doi.org/10.1007/s10758-019-09408-7

ORIGINAL RESEARCH

Factors Affecting Students’ Performance in Higher Education:

A Systematic Review of Predictive Data Mining Techniques

Amjed Abu Saa1 · Mostafa Al‑Emran2   · Khaled Shaalan1

Published online: 25 April 2019

© Springer Nature B.V. 2019

Abstract
Predicting the students’ performance has become a challenging task due to the increas-
ing amount of data in educational systems. In keeping with this, identifying the factors
affecting the students’ performance in higher education, especially by using predictive data
mining techniques, is still in short supply. This field of research is usually identified as
educational data mining. Hence, the main aim of this study is to identify the most com-
monly studied factors that affect the students’ performance, as well as, the most common
data mining techniques applied to identify these factors. In this study, 36 research articles
out of a total of 420 from 2009 to 2018 were critically reviewed and analyzed by applying
a systematic literature review approach. The results showed that the most common fac-
tors are grouped under four main categories, namely students’ previous grades and class
performance, students’ e-Learning activity, students’ demographics, and students’ social
information. Additionally, the results also indicated that the most common data mining
techniques used to predict and classify students’ factors are decision trees, Naïve Bayes
classifiers, and artificial neural networks.

Keywords  Educational data mining · Students’ performance · Data mining techniques ·

Systematic review

1 Introduction

An increasing interest has arisen during the past decade to identify the most important fac-
tors influencing students’ performance in higher education, especially by using data min-
ing methods and techniques. This field of research is usually identified as educational data

* Mostafa Al‑Emran
[email protected]
Amjed Abu Saa
[email protected]
Khaled Shaalan
[email protected]
1
Faculty of Engineering and IT, The British University in Dubai, Dubai, UAE
2
Applied Computational Civil and Structural Engineering Research Group, Faculty of Civil
Engineering, Ton Duc Thang University, Ho Chi Minh City, Vietnam

13
Vol.:(0123456789)
 568 A. Abu Saa et al.

mining (EDM) (Bakhshinategh et al. 2018). The motivation behind this interest is attrib-
uted to the applicability of such research in helping to identify low performing students
early enough to overcome their difficulties in learning and improve their learning out-
comes, which in turn serves the institutional goals of providing high-quality education eco-
systems. In addition, EDM is fast becoming an important field of research due to its ability
to extract new knowledge from a huge amount of students’ data (Wook et al. 2017). This
paper is equally interested in this topic, and our objective is to explore and review papers
from the past decade that are in the context of educational data mining and identifies the
main factors influencing students’ performance in higher education. EDM is defined by
the Educational Data Mining community website (www.educa​tiona​ldata​minin​g.org) as
“an emerging discipline, concerned with developing methods for exploring the unique and
increasingly large-scale data that come from educational settings and using those methods
to better understand students, and the settings which they learn in”. In general, EDM is a
set of methods that apply data mining techniques such clustering, classification, prediction
among the others to the data retrieved by many educational systems (Berland et al. 2014).
Researchers in this field tend to study different types of students’ factors and attributes
that affect students’ performance and learning outcomes (Abu Saa et  al. 2019). Shahiri
et al. (2015) conducted a systematic literature review on predicting students’ performance
using data mining techniques. The review tackled many subjects, one of which was to iden-
tify the important attributes used in predicting the students’ performance. Results showed
that the cumulative grade point average and internal assessments are the most frequent
attributes used for predicting the students’ performance. Furthermore, other important
attributes were also identified, including students’ demographic and external assessments,
extra-curricular activities, high school background, and social interaction network. Addi-
tionally, the results showed that the Decision tree and Neural Networks were the most fre-
quently used data mining techniques for predicting students’ performance. Besides, Peña-
Ayala (2014) conducted a survey and meta-analysis of research studies related to EDM.
The results indicated that 60% of EDM research articles have used predictive data mining
approaches as opposed to 40% which have used the descriptive approaches. Furthermore,
the results also indicated that classification and clustering were the most typical techniques
used by EDM research. Additionally, Bayes theorem, decision trees, instance-based learn-
ing (IBL), and hidden Markov model (HMM) were found to be the most popular meth-
ods used by EDM research. Furthermore, Romero and Ventura (2007) carried out a review
study aiming to analyze the application of data mining for different educational systems:
traditional system, web-based courses, content management systems, and intelligent web-
based systems. The results suggested investigating the applicability of using data mining
techniques for e-learning systems.
Predicting the students’ performance has become a challenging task due to the increas-
ing amount of data in educational systems (Shahiri et al. 2015). It is also argued that the
existing prediction methods are still insufficient to determine the appropriate techniques for
predicting the students’ performance in higher educational institutions. Moreover, the iden-
tification of the factors affecting the students’ performance is still neglecting and requires
further research. Thus, there is a clear need to identify the most important factors that were
found significant and truly affect the students’ performance from the increasing amount of
EDM literature. The aim of this paper is to investigate the literature related to EDM and to
identify the most important and most studied factors influencing the students’ performance
in higher education, as well as, to produce a generalized set of factors and attributes that
are believed to affect the students’ performance and learning outcomes in the higher educa-
tion sector. In order to perform this study, the existing literature has been reviewed using

13
Factors Affecting Students’ Performance in Higher Education:… 569

a systematic literature review (SLR) method. SLR is one of the most common approaches
used for literature review (Al-Emran et  al. 2018), and it serves our objective which sup-
posed to provide a summary of studies related to EDM and identify the factors affecting
students’ performance in higher education.
The rest of the paper is organized as follows. The methodology of the systematic litera-
ture review is explained in the next section. Section  3 demonstrates the results achieved
from this study. Discussing the results is outlined in Sect.  4, whereas the concluding
remarks are demonstrated in Sect. 5.

2 Systematic Review Method

In this paper, we employed a standard SLR methodology, which follows the guidelines
proposed by Kitchenham et al. (2009). SLR has many advantages over simple and unstruc-
tured literature review methods, as it is more likely to be considered reliable and unbiased
(Al-Qaysi et al. 2018). Additionally, information gathered from SLRs is highly reliable as
it was derived from various sources (Kitchenham and Charters 2007). There are three main
phases of SLR, namely planning, conducting, and reporting (Kitchenham et al. 2009). The
first two phases are discussed in the following two sub-sections, whereas the third phase is
discussed in Sect. 3.

2.1 Planning

Planning represents the first phase of the SLR method which includes five steps as per the
following sub-sections.

2.1.1 Identify the Research Goal and Research Questions

Our objective in this paper is to systematically review relevant literature through the SLR
process (Kitchenham and Charters 2007), and our research questions are provided as
follows:

• What are the factors affecting students’ performance in higher education?

• What are the data mining techniques used to analyze and predict the students’ perfor-
mance?

2.1.2 Identify the Keywords

Our search keywords were mostly driven by the research questions stated in the previous
subsection. After identifying the search keywords, we had to prepare a search string that
should work with the search engines of the libraries to be searched which will be identified
in the next section. The search string used in this study is: [(“data mining” OR “educa-
tional data mining”) AND (“factors affecting students’ performance” OR “predicting stu-
dents’ performance”)].
As it can be seen in the search string, the term “predicting students’ performance” was
added to the search string even though it did not appear in the research questions. This
is because we have noticed in the planning stage that there are a lot of research studies
that included this term in their titles and/or abstracts which indicates that the research is

13
 570 A. Abu Saa et al.

related to EDM, and these research studies predict the students’ performance based on
other attributes.

2.1.3 Identify the Sources

The following online library databases and search engines were selected to be searched for
the current SLR: ScienceDirect, EBSCO, ProQuest, JSTOR, and Taylor & Francis. The
authors of this study assumed that these databases are the main sources for collecting arti-
cles related to EDM and students’ performance.

2.1.4 Identify the Inclusion/Exclusion Criteria

Our inclusion criteria are shown in Table 1. Each study found in the search results must
meet these criteria in order to be included in our SLR.

2.1.5 Identify the Data Extraction Strategy

In this study, the data were collected based on the fields described in Table 2. The research
articles that did not have one or more fields from the data described in Table  2 were
excluded from the study.

2.2 Conducting the Review

Conducting the review represents the second phase of the SLR method which includes five
steps as per the following sub-sections.

2.2.1 Identify the Research

In this step, we started searching the online libraries’ databases with the aforementioned
search string. The initial search results returned by the search engines are illustrated in
Table 3.

2.2.2 Select the Studies

In this study, the selection of articles was carried out according to the Preferred Report-
ing Items for Systematic Reviews and Meta-Analysis (PRISMA) (Liberati et  al. 2009).

Table 1  Inclusion criteria Inclusion criteria

a. Must meet the research keywords conditions

b. Must be classified as a data mining or machine learning research
c. Must include the studied factors
d. Full-text articles must be available and accessible, and must not be
accessible via arXiv
e. Must not be a review paper
f. Must be published in the last decade (i.e., between 2009 and 2018)
g. Must be written in English language

13
Factors Affecting Students’ Performance in Higher Education:… 571

Table 2  Data layout
Item Item description

Paper ID An ID number is assigned to each research paper in order to be easily refer-

enced during the review
Source The database source of the research paper
Paper title The title of the research paper
Journal The journal that published the research paper
Author The author(s) of the research paper
Year The paper publication year
Country of study The country in which the study of the research paper was undertaken
Studied factors The list of factors that were studied in the research paper
Factors category(s) The categories of the factors in the previous field, such as: students demo-
graphics, students social information, e-Learning activities, etc.
Factors found significant The list of factors that were found significant by the researchers of this study
out of the full list of studied factors
Data mining approaches The data mining techniques used in the research paper, such as: classification,
clustering, etc.
Data mining algorithms The data mining algorithms that were used in the research paper, such as:
decision trees, SVM, K-Means clustering, etc.
Data collection techniques The techniques that were used to collect data in the research paper, such as:
surveys, student information systems data, e-Learning system data, etc.
Dataset size The size of the dataset that was used in the research paper

Table 3  Initial search results Online library database/search engine Results

Science direct 48
EBSCO 80
ProQuest 34
JSTOR 201
Taylor and Francis online 57
Total 420

PRISMA exhibits the flow of information throughout the SLR phases, in which it maps
the number of articles identified, included, excluded, and the reasons behind exclusions
(Liberati et  al. 2009). Figure  1 illustrates the PRISMA flowchart. In that, we select the
research articles that satisfy the inclusion/exclusion criteria and examine the contents of
the selected articles to verify their eligibility for selection. We then applied the automatic
and semi-automatic article selection. The process of automatic selection involves exam-
ining the articles against the titles, abstracts, and keywords, whereas the semi-automatic
selection involves reading the full-text of the remaining articles.
As a result of the automatic selection, we found 218 duplicate articles, in which, 57 did
not meet the research keywords and were not related to the subject of this SLR study, 19
were not a data mining research, 23 were not free access nor available, or an arXiv paper,
21 were review papers, and 34 were non-English papers. Furthermore, research dates were
set as a search filter in the search engine and were excluded from the initial search results.

13
 572 A. Abu Saa et al.

Fig. 1  PRISMA flowchart

Finally, we were left with 48 research papers. By applying the semi-automatic paper selec-
tion and reading the full-text articles, it was found that 12 papers did not include the factors
that affect the students’ performance; hence, they were also excluded from the list. There-
fore, the final dataset size of the SLR is 36 research papers. Table 11 (See “Appendix”) lists
all the 36 research papers selected for this SLR.

2.2.3 Assess the Study Quality

In order to assess the quality of the selected papers in the previous sub-section, we have to
answer the questions in Table 4 for each paper in the dataset. It is worth mentioning that
the analysis of the collected studies was carried out by the first and second authors of this
study by analyzing each article manually. The intercoder agreement rate for coding was
97.5%. The analysis differences between the two authors were resolved through discussion
and further review of the disputed studies.
Correspondingly, the answers to the questions in Table 4 accept three scores: Yes (1),
Partially (0.5), and No (0). Summing up the scores of all questions for each study will
result in a cumulative quality score for each paper out of 9. The result is then converted

13
Factors Affecting Students’ Performance in Higher Education:… 573

Table 4  Quality assessment # Question

questions (Kitchenham and
Charters 2007)
Q1 Are the study aims clearly stated?
Q2 Is the research described adequately?
Q3 Does the study explore diversity of perspectives and contexts?
Q4 Do the objectives lead to conclusions clearly?
Q5 Are the findings important?
Q6 Are negative findings presented?
Q7 Do the researchers explain the consequences of any problems?
Q8 Does the study add to your knowledge or understanding?
Q9 Do the results add to the literature?

into a percentage, e.g., 7 out of 9 is 77.78%. Table 12 (See “Appendix”) shows the cumu-
lative quality scores for all the papers included in our dataset along with their percentage
results, sorted by percentage.
As shown in Table  12, 34 out of 36 papers achieved a score equal to or greater than
4.5 (50%), whereas only two scored a value of 4/9 (44.44%) due to their low quality and
poor content. The most top scored papers are RP10 and RP22 with a score of 8/9 (88.9%).
Consequently, the two low-scored papers (RP130, RP136) were removed from the SLR
process, and the remaining 34 were kept for the subsequent steps.

2.2.4 Extract the Data

In this step, the required data for our SLR study will be extracted from the selected papers
according to the data layout in Table 2. We focused in this step on finding the factors that
affect the students’ performance, and most importantly, those that were found significant
by the researchers in their papers, as well as, the data mining techniques and algorithms
used by the researchers in their data mining research. These data will help us to get useful
insights and results that will empower us to answer our research questions. Table 13 (See
“Appendix”) summarizes the extracted data for each paper in our dataset.

2.2.5 Synthesize the Data

We extracted 215 distinct significant factors from the 34 research papers that affect the
performance of students in their education life. Furthermore, during the data extraction
process, we identified the category for each set of factors that were collected from each
article. As a result, we found nine-factor categories that the 215 factors belong to. Table 5
shows an extract of the nine categories of factors along with their descriptions, whereas the
sources and number of articles for these categories are illustrated in the extended part of
the Table 14 in “Appendix”.

3 Results

In this section, we report the results of our SLR study, where we will answer our research
questions, and elaborate on the interesting results we came up with from the extracted data.

13
 574 A. Abu Saa et al.

Table 5  Description of factors’ categories (extract)

Category Description

Students e-Learning activity The activity logs of students in e-Learning systems, such as, the number
of logins, the number of assignments done, number of quizzes done,
etc.
Students previous grades and The grades or other performance indicators of students in previous
class performance courses, semesters, or years
Students environment The attributes of a student environment, such as: the type of school, the
type of classroom, class period, etc.
Students demographics The demographic data of a student, such as: gender, age, nationality,
ethnicity, etc.
Instructor attributes Information about the instructor of the student and his/her evaluation
results
Course attributes Information about the course or module the student is taking, such as,
length of the course, difficulty, etc.
Students social information Information related to the student social life, like the number of friends,
if s/he smokes or not, etc.
Course evaluations Data collected from course evaluation surveys, such as, questions related
to the clarity of the course, the level of satisfaction, etc.
Students experience information Information about the experience of students about the course, such as
the readiness of the student, and self-efficacy

3.1 Research Question 1: Factors Affecting Students’ Performance

3.1.1 Distribution of Research Articles by Factors’ Categories

We classified each article under one or more categories as described in Table  5. These
include: (1) Students e-Learning activity, (2) Students previous grades and class perfor-
mance, (3) Students environment, (4) Students demographics, (5) Instructor attributes, (6)
Course attributes, (7) Students social information, (8) Course evaluations, and (9) Students
experience information. As it can be seen in Fig.  2, the most common and widely used

Fig. 2  Distribution of research articles by category

13
Factors Affecting Students’ Performance in Higher Education:… 575

factor categories for predicting students’ performance in higher education are students’
previous grades and class performance (26%), followed by students’ e-Learning activity
(25%), students’ demographics (23%), and students’ social information (12%), respectively.
These 4 categories were presented in 86% of the analyzed research studies.
This finding is in agreement with the findings of a prior systematic literature review
conducted by Shahiri et al. (2015), which showed that the CGPA and internal assessment
marks are the most frequently used attributes in the EDM community for predicting the
students’ performance. This matches our top factor category which represents the “stu-
dents’ previous grades and class performance”. The other five categories which represent a
total of 14% altogether were used by a few number of studies and did not frequently appear
in other research articles; therefore, they were considered as ad hoc factors.

3.1.2 Distribution of Research Articles by Publication Year

Table  6 indicates that educational data mining research field was most popular in 2016,
where more than 17.5% of the research was conducted in this year, and a significant
increase of interest was started on 2012.

3.1.3 Distribution of Research Articles by Data Collection Techniques

Another dimension was added to the data collection, which is the techniques of the data
collection. In fact, we identified five techniques of data collection, namely: (1) e-Learning
system logs, (2) Student information system data, (3) Surveys, (4) Course evaluations sur-
veys, and (5) Network access logs. Table 7 summarizes the data collection techniques used
by research papers in our dataset.

3.2 Research Question 2: Data Mining Techniques Used to Analyze and Predict

Students’ Performance

The main data mining approaches used in most of the analyzed studies are: (1) classifi-
cation and (2) clustering. Table  8 summarizes the distribution of research articles in our
dataset by the two data mining approaches described. The main data mining approach used

Table 6  Distribution of research Year Papers Number

articles by publication year of papers

2009 RP142 1
2010 RP28, RP35, RP113 3
2011 RP120, RP127, RP128 3
2012 RP23, RP25, RP33, RP43 4
2013 RP16, RP34, RP323 3
2014 RP9, RP44, RP67, RP123 4
2015 RP1, RP81, RP174, RP220 4
2016 RP10, RP12, RP22, RP36, RP75, 6
RP198
2017 RP2, RP7, RP69, RP94 4
2018 RP5, RP277 2

13
 576

13
Table 7  Summary of data collection techniques
Data collection technique Description Articles No. of articles

e-Learning system logs Logs obtained from e-Learning systems RP1, RP7, RP9, RP23, RP25, RP33, RP34, RP36, RP44, 14
RP94, RP113, RP198, RP220, RP277
Student information system data Extracted data from student information systems, such as RP2, RP5, RP16, RP22, RP25, RP28, RP43, RP120, 11
demographic data, admission data, grades, etc. RP123, RP127, RP323
Surveys General survey obtained from students directly RP16, RP22, RP67, RP69, RP75, RP128, RP142 7
Course evaluations surveys Answers of course evaluation surveys, generally obtained at RP10, RP12, RP25, RP35 4
the end of each course
Network access logs Network dumb logs of students’ activities on the internet RP174 1
and the university network
A. Abu Saa et al.
Factors Affecting Students’ Performance in Higher Education:… 577

Table 8  Distribution of Data mining approach Papers Num-

research articles by data mining ber of
approaches papers

Classification All 34
Clustering RP2, RP23, RP25, 4
RP34

is classification. It was found that all the research papers in the dataset have used the classi-
fication approach to classify and predict the students’ performance. On the other hand, only
4 research papers have used the clustering along with classification which was useful in
order to find out how many different groups of students available in the dataset and extract
specific features of each group. This finding is in line with the findings of a prior study car-
ried out by Peña-Ayala (2014), where it showed that the classification and clustering were
the most typical data mining techniques used by EDM research.
Furthermore, we extracted 141 data mining techniques/algorithms used by the 34 papers
in our dataset. Out of which, 74 were distinct. The algorithms are: 1NN, 3NN, ADTree,
Apriori Algorithm, Artificial Neural Networks, BayesNet, Bivariate Regression, BP, C4.5
Decision Tree, CART Decision Tree, CHAID, CitationKNN, Clustering, CRT Decision
Tree, Decision Tree (DT), DecisionStump, DTNB, EM, FarthestFirst, Feed-Forward Neural
Network (FFNN), G3P-MI, GP-ICRM, Gradient Boosting (GBM), HierarchicalClusterer,
IBk, ICRM v1, ICRM v2, ICRM v3, ICRM2, ID3 Decision Tree, J48, Jrip, K-Means Clus-
tering, K-Nearest Neighbour (k-NN), LADTree, LGR, Locally weighted linear regression,
Logistic Regression, MILR, Multi-layer Perceptron (MLP) neural network, Model Trees,
Multi-logistic Regression (MLR), Naïve Bayes classifiers, NaiveBayesSimple, Neural Net-
work (NN), NLPCA, Nnge, OneR, PART, Prism, Probabilistic ensemble SFAM classifier
(PESFAM), Probabilistic Ensemble Simplified Fuzzy ARTMAP, Proportional Odd Model
(POM), Radial Basis Function (RBF) Network, Random Forest, Random Tree, RBF Net-
work, Regression, Regression neural network model (RNN), REPTree, Resonance Theory
Mapping (PESFAM), Ridor, RIPPER, Rule Induction, sIB, SimpleCart, SimpleKMeans,
SMO, Support Vector Machine (SVM), Support Vector Ordinal Regression (SVOR), Sys-
tem for Educational Data Mining (SEDM), Visualization, WINNOW, Xmeans. However,
the most commonly used algorithms that were used in 4 or more research papers (i.e., in
more than 10% of the papers), are shown in Table 9.
Furthermore, we merged similar algorithms together into one category, for example,
ID3 and C4.5 are both decision trees, so we grouped them together under the decision tree
category. After doing so, we ended up with 7 categories of algorithms. Table 10 shows the
7 groups of algorithms and the frequency of their usage in the analyzed research articles.
As it can be observed from Table 10, the most commonly used categories of data mining
algorithms are Decision Trees, Naïve Bayes classifiers, and Artificial Neural Networks.

4 Discussion

This study reports a systematic literature review regarding the students’ academic perfor-
mance in higher education. The study was designed to identify the most commonly studied
factors that affect the students’ performance, as well as, the most common data mining
techniques applied to identify these factors. The study reviewed 34 research articles related

13
 578 A. Abu Saa et al.

Table 9  Most commonly used Algorithm Frequency Percentagea (%)

data mining algorithms
Naïve Bayes classifiers 13 38.2
Support vector machine (SVM) 8 23.5
Logistic regression 6 17.6
K-Nearest neighbor (k-NN) 5 14.7
ID3 Decision tree 4 11.8
C4.5 Decision tree 4 11.8
Decision tree (DT) 4 11.8
Multi-layer perceptron (MLP) 4 11.8
neural network
Neural network (NN) 4 11.8
a
 The percentage calculated is out of the total number of analyzed arti-
cles (N = 34)

Table 10  Most commonly used Algorithm Frequency Percentagea (%)

algorithms by category
Decision trees 35 24.8
Naïve Bayes classifiers 14 9.9
Artificial neural networks 13 9.2
Regression 12 8.5
Support vector machine 9 6.4
K-Nearest neighbor 8 5.7
K-Means 3 2.1
Other algorithms 47 33.3
a
 The percentage calculated is out of the total number of algorithms
(N = 141)

to the subject and came up with results of research distribution across multiple dimensions.
In terms of the first research direction, the main results showed that students’ previous
grades and class performance, students’ e-Learning activity, students’ demographics, and
students’ social information are generally the main factors affecting the students’ academic
performance in higher education.
According to the literature (Asif et al. 2017; Burgos et al. 2018; Gómez-Rey et al. 2016;
Kotsiantis et  al. 2010; Márquez-Vera et  al. 2013, 2016), the first category which relates
to the students’ previous grades and class performance was also regarded as one of the
influential factors that could affect the students’ academic performance. This finding could
be explained that students’ performance remains equally the same across their educational
life. In other words, if the student is in the habit of having good grades at the beginning
of his/her studies, he/she would remain good for the rest of his/her academic life. This is
equally the same for students who tend to score bad grades, they might also keep the same
pattern across their study, and definitely, this could affect their performance in the cur-
rent and future studies. For the higher educational institutions, these results could assist the
education stakeholders to focus on the specific areas of weaknesses in the students’ aca-
demic life and try to overcome these shortcomings by improving the students’ educational
outcomes and quality.

13
Factors Affecting Students’ Performance in Higher Education:… 579

The results were also in agreement with the existing literature (Abdous et al. 2012; Bur-
gos et  al. 2018; Hung et  al. 2012; Lara et  al. 2014; Xing et  al. 2015; Zafra and Ventura
2012), in which the second category of factors (i.e., students’ e-Learning activity) was also
found to have a significant impact on students’ academic performance. This indicates that
the more the students engage in e-Learning activities (e.g., accessing online material, solv-
ing online quizzes, and uploading assignments into the e-Learning system), the more likely
the students achieve higher grades and improve their overall performance. Practically,
these results could assist the educational institutions to concentrate on students’ e-Learning
activities and promote using e-Learning systems in order to increase the students’ perfor-
mance and education quality (Salloum et al. 2019).
Concerning the other two categories of factors which relate to the students’ demograph-
ics and students’ social information, it is posited that these categories are more students’
specific categories that deal with the students’ background and their surroundings behav-
iors. These two factors mainly depend on the students themselves, and that the students
should take care of such factors and try to avoid any behaviors or social activities that
might affect their academic performance. Moreover, the policy-makers can utilize such fac-
tors to create focus groups to take care of the students and provide them with special atten-
tion during their study in the institution.
With respect to the data collection techniques, the results pointed out that e-Learning
system logs and student information system data were the most frequent data collection
techniques used in the analyzed studies. Further research should consider other data collec-
tion techniques like surveys, course evaluation surveys, and network access logs as these
techniques were less used in the existing literature.
In terms of the second research direction, the results pointed out that Decision Trees,
Naïve Bayes classifiers, and Artificial Neural Networks are the most commonly used data
mining algorithms. This finding comes in line with the findings of prior systematic litera-
ture review studies (Peña-Ayala 2014; Shahiri et al. 2015), where it was found that decision
trees and Naïve Bayes classifiers are the most frequent data mining techniques used among
EDM research. Given these results at hand, further research is suggested to refer to other
data mining algorithms which might add more significant and reasonable conclusions.
Although the current systematic review partially shares the same research questions as
with Shahiri et  al. (2015) and Peña-Ayala (2014), it also has several differences that we
need to discuss. First, the time span of the analyzed research articles ranges between 2002
and 2015 in terms of the review study conducted by Shahiri et  al. (2015), and between
2010 and 2013 in terms of the review study carried out by Peña-Ayala (2014). In compari-
son with these reviews, the time span of the analyzed studies in this review study ranges
between 2009 and 2018. Second, with regard to the search strategy, Shahiri et al. (2015) has
mainly focused on collecting studies related to “students’ performance” and “educational
data mining”, whereas Peña-Ayala (2014) has concentrated on collecting studies related to
“EDM approaches” and “EDM tools”. To make this review study more distinctive, it has
focused on collecting studies related to “factors affecting students’ performance” and “data
mining or educational data mining techniques”. Changing the search strategy allows us to
retrieve different articles from those that were retrieved in previous reviews.
Third, Shahiri et al. (2015) concluded that Decision tree and Neural Networks were the
most commonly used data mining techniques, whereas Peña-Ayala (2014) indicated that
Bayes theorem, Decision tree, instance-based learning (IBL), and hidden Markov model
(HMM) were the most frequent techniques used. In this study, the results were partially
supported, in which Decision tree, Naïve Bayes classifiers, and Artificial Neural Networks
were observed to be the most common data mining techniques used to predict students’

13
 580 A. Abu Saa et al.

performance. In general, the differences between this review study and the previous ones
lie in the gap between time span, search strategy, and results achieved. Thus, this study
could serve as a comprehensive reference for pursuing further research in EDM in general,
and students’ performance in particular.

5 Conclusion and Future Work

In this paper, we identified the most common and widely studied factors that affect stu-
dents’ performance in higher education, as well as, the most common data mining
approaches, techniques, and algorithms used to classify and predict students’ performance.
We followed a systematic literature review methodology that consisted of multiple phases
and steps. This process is started by planning the review, from forming up the research
questions, through setting up the inclusion and exclusion criteria, until deciding the data
extraction strategy. Furthermore, the second phase consisted of the steps for conducting
the review, kicked off by searching and identifying the research papers for the literature
review, passing by assessing the quality of the selected research papers, and ended by
extracting the data and synthesize it. Finally, we ended up by reporting the results of our
SLR research, which concluded that the most common and widely used factors for pre-
dicting students’ performance in higher education are students’ previous grades and class
performance, students’ e-Learning activity, students’ demographics, and students’ social
information. Additionally, the results also showed that the most common and widely used
data mining techniques in the EDM field are Decision Trees, Naïve Bayes classifiers, and
Artificial Neural Networks.
As a future work, researchers can benefit from the outcomes of this systematic literature
review by employing it to their future research, particularly, the main results that highlight
the most frequently used categories of factors affecting students’ performance, as well as,
the most frequently used data mining techniques. Not to mention, having a generic set of
factors’ categories provides several possibilities to tailor the use of these categories and
come up with specific factors within the category of each educational institution, since it
might differ from one place to another, and from time to time. As a limitation, this study
has focused on some databases in terms of articles collection. In addition, future attempts
may consider other databases and search engines for articles collection in order to leverage
the number of analyzed studies.

Appendix

See Tables 11, 12, 13 and 14.

13
 Table 11  Selected research papers
Paper ID Source Journal Author

RP1 ScienceDirect Computers in Human Behavior Xing et al. (2015)

RP2 ScienceDirect Computers and Education Asif et al. (2017)
RP5 ScienceDirect Journal of Business Research Fernandes et al. (2018)
RP7 ScienceDirect Computers and Electrical Engineering Burgos et al. (2018)
RP9 ScienceDirect Computers and Education Lara et al. (2014)
RP10 EBSCO Expert Systems Gómez-Rey et al. (2016)
RP12 ProQuest Informatics in Education Jiang et al. (2016)
RP16 ProQuest Applied Intelligence Márquez-Vera et al. (2013)
RP22 EBSCO Expert Systems Márquez-Vera et al. (2016)
RP23 JSTOR Journal of Educational Technology and Society Abdous et al. (2012)
RP25 JSTOR Journal of Educational Technology and Society Hung et al. (2012)
RP28 ScienceDirect Knowledge-Based Systems Kotsiantis et al. (2010)
RP33 ScienceDirect Applied Soft Computing Zafra and Ventura (2012)
Factors Affecting Students’ Performance in Higher Education:…

RP34 ScienceDirect Computers and Education Romero et al. (2013)

RP35 EBSCO Applied Stochastic Models in Business and Industry Costantini et al. (2010)
RP36 EBSCO Expert Systems Gamulin et al. (2016)
RP43 ProQuest The Artificial Intelligence Review Kotsiantis (2012)
RP44 ScienceDirect Computers in Human Behavior Hu et al. (2014)
RP67 JSTOR European Journal of Open, Distance and E-Learning Yukselturk et al. (2014)
RP69 JSTOR Journal of Computer Science Abazeed and Khder (2017)
RP75 Google Scholar International Journal of Advanced Computer Science and Applications Abu Saa (2016)
RP81 JSTOR Indian Journal of Science and Technology Anuradha Bharathiar (2015)
RP94 JSTOR The Electronic Journal of Information Systems in Developing Countries Mwalumbwe and Mtebe (2017)
RP113 ScienceDirect Computers and Education Macfadyen and Dawson (2010)
RP120 Google Scholar International Journal of Advanced Computer Science and Applications Baradwaj and Pal (2012)
581

13
 Table 11  (continued)
582

Paper ID Source Journal Author

13
RP123 Google Scholar World Journal of Computer Application and Technology Badr El Din Ahmed and Sayed Elaraby (2014)
RP127 Google Scholar International Journal of Computer Science and Information Technologies Pandey and Pal (2011)
RP128 Google Scholar International Journal of Computer Science and Information Security Bhardwaj and Pal (2012)
RP130 Google Scholar International Journal of Innovative Technology and Creative Engineering Yadav et al. (2012)
RP136 Google Scholar World of Computer Science and Information Technology Journal Yadav and Pal (2012)
RP142 ScienceDirect Computers and Education Araque et al. (2009)
RP174 ProQuest International Conference on Digital Information and Communication Technology Zhou et al. (2015)
and its Applications
RP198 ScienceDirect Computers and Education Cerezo et al. (2016)
RP220 ProQuest The International Journal of Information and Learning Technology Chamizo-Gonzalez et al. (2015)
RP277 EBSCO Journal of AI and Data Mining Hasheminejad and Sarvmili (2018)
RP323 ScienceDirect Computers and Education Huang and Fang (2013)
A. Abu Saa et al.
Factors Affecting Students’ Performance in Higher Education:… 583

Table 12  Quality scores and percentages

Paper ID Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Quality score Percentage (%)

RP10 1 1 1 1 1 0 1 1 1 8 88.89
RP22 1 1 1 1 1 0 1 1 1 8 88.89
RP113 1 1 1 1 1 0 0.5 1 1 7.5 83.33
RP1 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP2 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP5 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP7 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP9 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP12 1 1 1 1 1 0 0 1 1 7 77.78
RP23 1 1 1 1 1 0 0.5 1 0.5 7 77.78
RP28 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP33 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP34 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP44 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP75 1 1 1 1 1 0 0 1 1 7 77.78
RP142 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP174 1 1 1 1 1 0 0 1 1 7 77.78
RP198 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP323 1 1 1 1 1 0 0.5 0.5 1 7 77.78
RP220 1 1 1 1 0.5 0 0.5 0.5 1 6.5 72.22
RP16 1 1 1 1 0.5 0 0 0.5 1 6 66.67
RP25 1 1 1 1 0.5 0 0.5 0.5 0.5 6 66.67
RP123 1 0.5 0.5 1 1 0 0 1 1 6 66.67
RP35 1 1 1 1 0.5 0 0 0.5 0.5 5.5 61.11
RP36 1 1 1 1 0.5 0 0.5 0 0.5 5.5 61.11
RP43 1 1 1 1 0.5 0 0 0.5 0.5 5.5 61.11
RP67 1 1 1 0.5 0.5 0 0.5 0 1 5.5 61.11
RP69 1 1 1 0.5 1 0 0 0 1 5.5 61.11
RP94 1 0.5 0.5 1 0.5 0 1 0 1 5.5 61.11
RP127 1 0.5 0.5 1 1 0 0 0.5 1 5.5 61.11
RP277 1 1 1 1 0.5 0 0 0.5 0.5 5.5 61.11
RP120 1 0.5 0.5 0.5 0 0 1 1 0.5 5 55.56
RP81 1 0 0.5 1 0.5 0 0.5 0.5 0.5 4.5 50.00
RP128 1 0.5 0.5 0.5 1 0 0 0.5 0.5 4.5 50.00
RP130 1 0.5 0.5 0.5 0.5 0 0 0.5 0.5 4 44.44
RP136 1 0.5 0.5 0.5 0.5 0 0 0.5 0.5 4 44.44

13
 Table 13  Summary of factors influencing students’ performance and used data mining approaches and techniques
584

Paper ID Factors category Factors found significant Data mining approach Data type(s)

13
RP1 Students e-Learning activity Chat logs of all messages that students send to Classification e-Learning system logs
each other in the group
Awareness records of actions of erasing the chat
messages on the chat bar
Geogebra logs of information on how students
virtually construct a geometry artifact (adding a
point or updating a segment)
System logs of students joining a virtual room,
leaves a virtual room or views different tabs
WhiteBoard logs of more specific actions on how
tools are being used in the white board areas
such as resizing objects or creating a textbox
RP2 Students previous grades and class performance High school marks (total and subject specific) 1. Classification Admission data
First and second year university courses’ marks 2. Clustering
RP5 1. Students environment Grades for the first 2 months Classification Student information system data
2. Students demographics Student’s place of residence—neighborhood
3. Students previous grades and class perfor- School name
mance School subjects
Absences
Student’s place of residence—city
Age
RP7 Students e-Learning activity 12 Assessment activities from e-Learning system Classification e-Learning system logs
Teaching schedule
RP9 Students e-Learning activity Number of virtual classroom accesses by the Classification e-Learning system logs
student in the week in question
Number of different days of the week on which
the student accesses the virtual classroom
Whether or not the resource has been visualized
in the week in question
Number of times that the student has visualized
the resource in the week in question
A. Abu Saa et al.
 Table 13  (continued)
Paper ID Factors category Factors found significant Data mining approach Data type(s)

RP10 1. Instructor attributes Instructor’s knowledge Classification Course evaluations surveys

2. Students previous grades and class perfor- Instructor’s effective use of the class hours
mance Instructor’s coherence with lesson plan
3. Course attributes Openness and respect of the instructor to stu-
dents’ views
Instructor’s positive approach to students
Instructor readiness for classes
Instructor explanations about the course and
instructor helpfulness
RP12 1. Instructor attributes Instructor’s organization and clarity Classification Course evaluations surveys
2. Course attributes Instructor’s response to questions
Instructor’s visual presentation
Instructor’s encouragement to think indepen-
dently
Instructor’s attitude towards teaching
Professor-class relationship
Factors Affecting Students’ Performance in Higher Education:…

Difficulty of concepts covered

Contribution of assignments to understanding of
concepts
How well tests reflect the course material
Attendance (the number of evaluations received
divided by course enrolment)
585

13
 Table 13  (continued)
586

Paper ID Factors category Factors found significant Data mining approach Data type(s)

13
RP16 1. Students previous grades and class perfor- Scores in specific subjects Classification 1. Surveys
mance Level of motivation 2. Student information system data
2. Students demographics GPA in secondary school
3. Students social data Age
Number of brothers/sisters
Classroom/group
Smoking habits
Studying in group
Marital status
Time spent doing exercises
RP22 1. Students previous grades and class perfor- GPA in secondary school Classification 1. Surveys
mance Classroom/group enrolled 2. Student information system data
2. Students demographics Number of students in the group/class
3. Students social data Age
Attendance during morning/evening sessions
Having a job
Mother’s level of education
RP23 Students e-Learning activity Students activity data from an online video 1. Classification e-Learning system logs
e-learning system 2. Clustering
Number of questions
Number of chat messages
Total login times
Final grade
A. Abu Saa et al.
 Table 13  (continued)
Paper ID Factors category Factors found significant Data mining approach Data type(s)

RP25 1. Students e-learning activity 1. Students e-Learning activity 1. Classification 1. e-Learning system logs
2. Students demographics Average frequency of logins per course 2. Clustering 2. Student information system data
3. Course evaluations Average frequency of tab accessed per course 3. Course evaluations surveys
Average frequency of module accessed per course
Average frequency of clicks per course
Average frequency of course accessed per course
Average frequency of page accessed per course
Average frequency of course content accessed
per course
Average number of discussion board entries per
course
2. Students demographics: age, gender, gradua-
tion year, city, school district, number of online
course(s) taken, number of online course(s)
passed, number of online course(s) failed, and
final grade average
3. Student information: number of courses taken,
Factors Affecting Students’ Performance in Higher Education:…

number of courses failed, number of courses

passed, average individual student pass rate for
all courses in academic year 2009–2010
RP28 Students previous grades and class performance 1st written assignment Classification Student information system data
2nd written assignment
3rd written assignment
4th written assignment
587

13
 Table 13  (continued)
588

Paper ID Factors category Factors found significant Data mining approach Data type(s)

13
RP33 Students e-Learning activity Number of pieces of coursework done by the user Classification e-Learning system logs
in the course
Total time in seconds that the user has taken in
the assignment section
Number of messages sent by the user in the forum
Number of messages read by the user in the
forum
Total time in seconds that the user has taken in
the forum section
Number of quizzes seen by the user
Number of quizzes passed by the user
Number of quizzes failed by the user
Total time in seconds that the user has taken in
the quiz section
RP34 Students e-Learning activity Number of messages written by the student 1. Classification e-Learning system logs
Number of words written by the student 2. Clustering
Average score on the instructor’s evaluation of
the student’s messages
Degree centrality of the student
Degree prestige of the student
A. Abu Saa et al.
 Table 13  (continued)
Paper ID Factors category Factors found significant Data mining approach Data type(s)

RP35 Course evaluations Program workload Classification Course evaluations surveys

Program organization of teaching
Keep scheduled hours
Clear exam rules
Availability of lecturer outside class
Student’s previous knowledge of the topic
Teacher ability to motivate
Clarity of teaching
Availability of lecturer inside class
On schedule with program
Workload-credit ratio
Prescribed reading list
Adequacy of lecture hall
Student interest in topic
Overall class satisfaction
RP36 Students e-Learning activity Student access time series Classification e-Learning system logs
Factors Affecting Students’ Performance in Higher Education:…

Number of clicks per course.

RP43 1. Students previous grades and class perfor- 4th Written assignment Classification Student information system data
mance 3rd Written assignment
2. Students demographics
589

13
 Table 13  (continued)
590

Paper ID Factors category Factors found significant Data mining approach Data type(s)

13
RP44 Students e-Learning activity Total time online (s) Classification e-Learning system logs
Number of course material viewed (by material
category) (s)
Average time per session (s)
Total time material viewed (s)
Number of course material viewed
# Course material viewed (by material category)/
number of Course material released to date
Number of logins
Average time material viewed (s)
Total time course material viewed (by material
category) (s)
RP67 1. Students demographics Online learning readiness Classification Surveys
2. Students experience data Previous online experience
Gender
Online technologies self-efficacy
Age
Prior knowledge
RP69 1. Students demographics Gender Classification Surveys
2. Students previous grades and class perfor- High school grade
mance Major in high school
Previous GPA
Number of courses registered
Sponsor
Advisory visit
English score
Attendance
Core versus elective
Study time
Performance
A. Abu Saa et al.
 Table 13  (continued)
Paper ID Factors category Factors found significant Data mining approach Data type(s)

RP75 1. Students demographics Gender Classification Surveys

2. Students previous grades and class perfor- High school grade
mance Mother occupation status
3. Students social data Discount
RP81 1. Students demographics Previous semester marks Classification Not reported
2. Students previous grades and class perfor- Family annual income
mance Student category
3. Students social data Family size
Attendance
High school grade
Assignment performance
RP94 Students e-Learning activity Interactions with peers Classification e-Learning system logs
Number of exercises performed
Number of forum posts
RP113 Students e-Learning activity Total # discussion messages posted Classification e-Learning system logs
Total number of online sessions
Factors Affecting Students’ Performance in Higher Education:…

Total time online

# Files viewed
# Assessments finished
# Assessments started
# Reply discussion messages posted
# Mail messages sent
# Assignments submitted
# Discussion messages read
# Web links viewed
# New discussion messages posted
# Mail messages read
591

13
 Table 13  (continued)
592

Paper ID Factors category Factors found significant Data mining approach Data type(s)

13
RP120 Students previous grades and class performance Previous semester marks Classification Student information system data
Class test grade
Attendance
Assignment
Lab work
RP123 1. Students demographics Midterm marks Classification Student information system data
2. Students previous grades and class perfor- Lab test grades
mance Students practice
Homework
Seminar performance
High school branch
Attendance
RP127 Students demographics Gender Classification Student information system data
Language medium
Stream of bachelor’s degree
Division obtained
RP128 1. Students demographics Students grade in senior secondary education Classification Surveys
2. Students previous grades and class perfor- Living location
mance Medium of teaching
3. Students social data Mother’s qualification
Students other habit
Family annual income status
Students family status
RP142 1. Students demographics Father’s education level Classification Surveys
2. Students previous grades and class perfor- Mother’s education level
mance Academic performance rate
3. Students social data Average round
Mode round
Access year
A. Abu Saa et al.
 Table 13  (continued)
Paper ID Factors category Factors found significant Data mining approach Data type(s)

RP174 Students social data Number of records on each category of websites Classification Network access logs
Duration of watching online videos
Students’ grades on advanced mathematics
RP198 Students e-Learning activity Total time spent on practical tasks Clustering e-Learning system logs
The time taken to hand in the task since the task
was made available in the LE
Number of words in forum posts
RP220 Students e-Learning activity Assignment upload Classification e-Learning system logs
Forum add post
Forum update post
Forum view discussion
Assignment view
Assignment view all
Course view
Forum view forum
Resource view
Factors Affecting Students’ Performance in Higher Education:…

RP277 Students e-Learning activity Course identification number Classification e-Learning system logs
Number of assignments done
Number of quizzes passed
Number of quizzes failed
Number of messages send to forum
Number of messages read on the forums
Total time used on assignments
Total time used on quizzes
Total time used on forum
593

13
 Table 13  (continued)
594

Paper ID Factors category Factors found significant Data mining approach Data type(s)

13
RP323 Students previous grades and class performance Cumulative GPA Classification Student information system data
Statistics grade
Calculus I grade
Calculus II grade
Physics grade
Dynamics mid-exam #1 score
Dynamics mid-exam #2 score
Dynamics mid-exam #3 score
A. Abu Saa et al.
 Table 14  Description of factors’ categories (extended)
Category Papers Number
of articles

Students e-Learning activity RP1, RP7, RP9, RP23, RP25, RP33, RP34, RP36, RP44, RP94, RP113, RP198, RP220, RP277 14
Students Previous grades and class perfor- RP2, RP5, RP10, RP16, RP22, RP28, RP43, RP69, RP75, RP81, RP120, RP123, RP128, RP142, RP323 15
mance
Students environment RP5 1
Students demographics RP5, RP16, RP22, RP25, RP43, RP67, RP69, RP75, RP81, RP123, RP127, RP128, RP142 13
Factors Affecting Students’ Performance in Higher Education:…

Instructor attributes RP10, RP12 2

Course attributes RP10, RP12 2
Students social information RP16, RP22, RP75, RP81, RP128, RP142, RP174 7
Course evaluations RP25, RP35 2
Students experience information RP67 1
595

13
 596 A. Abu Saa et al.

References
Abazeed, A., & Khder, M. (2017). A classification and prediction model for student’s performance in uni-
versity level. Journal of Computer Science, 13, 228–233.
Abdous, M., He, W., & Yen, C. J. (2012). Using data mining for predicting relationships between online
question theme and final grade. Educational Technology and Society, 15(3), 77–88.
Abu Saa, A. (2016). Educational data mining and students’ performance prediction. Internation Journal of
Advanced Computer Science and Applications. https​://doi.org/10.14569​/IJACS​A.2016.07053​1.
Abu Saa, A., Al-Emran, M., & Shaalan, K. (2019). Mining student information system records to predict
students’ academic performance. In International conference on advanced machine learning tech-
nologies and applications (pp. 229–239). Berlin: Springer.
Al-Emran, M., Mezhuyev, V., Kamaludin, A., & Shaalan, K. (2018). The impact of knowledge manage-
ment processes on information systems: A systematic review. International Journal of Information
Management, 43, 173–187.
Al-Qaysi, N., Mohamad-Nordin, N., & Al-Emran, M. (2018). A systematic review of social media
acceptance from the perspective of educational and information systems theories and models. Jour-
nal of Educational Computing Research. https​://doi.org/10.1177/07356​33118​81787​9.
Anuradha Bharathiar, C., & Velmurugan, T. (2015). A comparative analysis on the evaluation of classi-
fication algorithms in the prediction of students performance. Indian Journal of Science and Tech-
nology. https​://doi.org/10.17485​/ijst/2015/v8i.
Araque, F., Rolddsán, C., & Salguero, A. (2009). Factors influencing university drop out rates. Comput-
ers and Education. https​://doi.org/10.1016/j.compe​du.2009.03.013.
Asif, R., Merceron, A., Ali, S. A., & Haider, N. G. (2017). Analyzing undergraduate students’ perfor-
mance using educational data mining. Computers and Education. https​://doi.org/10.1016/j.compe​
du.2017.05.007.
Badr El Din Ahmed, A., Sayed Elaraby, I., & Sayed Elaraby, I. (2014). Data mining: A prediction for
student’s performance using classification method. World Journal of Computer Application and
Technology. https​://doi.org/10.13189​/wjcat​.2014.02020​3.
Bakhshinategh, B., Zaiane, O. R., ElAtia, S., & Ipperciel, D. (2018). Educational data mining applica-
tions and tasks: A survey of the last 10 years. Education and Information Technologies. https​://doi.
org/10.1007/s1063​9-017-9616-z.
Baradwaj, B., & Pal, S. (2012). Mining educational data to analyze student’s performance. International
Journal of Advanced Computer Science and Applications, 2(6), 63–69.
Berland, M., Baker, R. S., & Blikstein, P. (2014). Educational data mining and learning analytics: Appli-
cations to constructionist research. Technology, Knowledge and Learning. https​://doi.org/10.1007/
s1075​8-014-9223-7.
Bhardwaj, B. K., & Pal, S. (2012). Data mining: A prediction for performance improvement using clas-
sification. (IJCSIS) International Journal of Computer Science and Information Security, 9(4), 1–5.
Burgos, C., Campanario, M. L., de la Peña, D., Lara, J. A., Lizcano, D., & Martínez, M. A. (2018). Data
mining for modeling students’ performance: A tutoring action plan to prevent academic dropout.
Computers and Electrical Engineering. https​://doi.org/10.1016/j.compe​lecen​g.2017.03.005.
Cerezo, R., Sánchez-Santillán, M., Paule-Ruiz, M. P., & Núñez, J. C. (2016). Students’ LMS interaction
patterns and their relationship with achievement: A case study in higher education. Computers and
Education. https​://doi.org/10.1016/j.compe​du.2016.02.006.
Chamizo-Gonzalez, J., Cano-Montero, E. I., Urquia-Grande, E., & Muñoz-Colomina, C. I. (2015). Edu-
cational data mining for improving learning outcomes in teaching accounting within higher educa-
tion. International Journal of Information and Learning Technology. https​://doi.org/10.1108/IJILT​
-08-2015-0020.
Costantini, P., Linting, M., & Porzio, G. C. (2010). Mining performance data through nonlinear PCA
with optimal scaling. Applied Stochastic Models in Business and Industry. https​://doi.org/10.1002/
asmb.771.
Fernandes, E., Holanda, M., Victorino, M., Borges, V., Carvalho, R., & Van Erven, G. (2018). Educa-
tional data mining: Predictive analysis of academic performance of public school students in the
capital of Brazil. Journal of Business Research. https​://doi.org/10.1016/j.jbusr​es.2018.02.012.
Gamulin, J., Gamulin, O., & Kermek, D. (2016). Using Fourier coefficients in time series analysis for
student performance prediction in blended learning environments. Expert Systems. https​://doi.
org/10.1111/exsy.12142​.
Gómez-Rey, P., Fernández-Navarro, F., & Barberà, E. (2016). Ordinal regression by a gravitational
model in the field of educational data mining. Expert Systems. https​://doi.org/10.1111/exsy.12138​.

13
Factors Affecting Students’ Performance in Higher Education:… 597

Hasheminejad, S. M., & Sarvmili, M. (2018). S3PSO: Students’ performance prediction based on parti-
cle swarm optimization. Journal of AI and Data Mining, 7, 77–96.
Hu, Y.-H., Lo, C.-L., & Shih, S.-P. (2014). Developing early warning systems to predict students’ online
learning performance. Computers in Human Behavior. https​://doi.org/10.1016/j.chb.2014.04.002.
Huang, S., & Fang, N. (2013). Predicting student academic performance in an engineering dynamics
course: A comparison of four types of predictive mathematical models. Computers and Education.
https​://doi.org/10.1016/j.compe​du.2012.08.015.
Hung, J., Hsu, Y.-C., & Rice, K. (2012). Integrating data mining in program evaluation of K-12 online
education. Educational Technology and Society. https​://doi.org/10.1207/s1532​7752j​pa850​2.
Jiang, Y. H., Javaad, S. S., & Golab, L. (2016). Data mining of undergraduate course evaluations. Infor-
matics in Education. https​://doi.org/10.15388​/infed​u.2016.05.
Kitchenham, B., & Charters, S. (2007). Guidelines for performing systematic literature reviews in soft-
ware engineering (pp. 1–57). Software Engineering Group, School of Computer Science and Math-
ematics, Keele University.
Kitchenham, B., Pearl Brereton, O., Budgen, D., Turner, M., Bailey, J., & Linkman, S. (2009). Sys-
tematic literature reviews in software engineering: A systematic literature review. Information and
Software Technology. https​://doi.org/10.1016/j.infso​f.2008.09.009.
Kotsiantis, S. B. (2012). Use of machine learning techniques for educational proposes: A decision sup-
port system for forecasting students’ grades. Artificial Intelligence Review. https​://doi.org/10.1007/
s1046​2-011-9234-x.
Kotsiantis, S., Patriarcheas, K., & Xenos, M. (2010). A combinational incremental ensemble of classi-
fiers as a technique for predicting students’ performance in distance education. Knowledge-Based
Systems. https​://doi.org/10.1016/j.knosy​s.2010.03.010.
Lara, J. A., Lizcano, D., Martínez, M. A., Pazos, J., & Riera, T. (2014). A system for knowledge discov-
ery in e-Learning environments within the European higher education area: Application to student
data from Open University of Madrid, UDIMA. Computers and Education, 72, 23–36. https​://doi.
org/10.1016/j.compe​du.2013.10.009.
Liberati, A., Altman, D. G., Tetzlaff, J., Mulrow, C., Gøtzsche, P. C., Ioannidis, J., et  al. (2009). The
PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate
health care interventions: Explanation and elaboration. Journal of Clinical Epidemiology. https​://
doi.org/10.1016/j.jclin​epi.2009.06.006.
Macfadyen, L. P., & Dawson, S. (2010). Mining LMS data to develop an “early warning system” for
educators: A proof of concept. Computers and Education. https​://doi.org/10.1016/j.compe​
du.2009.09.008.
Márquez-Vera, C., Cano, A., Romero, C., Noaman, A. Y. M., Mousa Fardoun, H., & Ventura, S. (2016).
Early dropout prediction using data mining: A case study with high school students. Expert Sys-
tems. https​://doi.org/10.1111/exsy.12135​.
Márquez-Vera, C., Cano, A., Romero, C., & Ventura, S. (2013). Predicting student failure at school using
genetic programming and different data mining approaches with high dimensional and imbalanced
data. Applied Intelligence. https​://doi.org/10.1007/s1048​9-012-0374-8.
Mwalumbwe, I., & Mtebe, J. S. (2017). Using learning analytics to predict students’ performance
in moodle learning management system: A case of Mbeya University of science and tech-
nology. Electronic Journal of Information Systems in Developing Countries. https​://doi.
org/10.1002/j.1681-4835.2017.tb005​77.x.
Pandey, U. K., & Pal, S. (2011). Data mining: A prediction of performer or underperformer using clas-
sification. International Journal of Computer Science and Information Technologies, 2, 686–690.
Peña-Ayala, A. (2014). Educational data mining: A survey and a data mining-based analysis of recent
works. Expert Systems with Applications. https​://doi.org/10.1016/j.eswa.2013.08.042.
Romero, C., López, M. A., Luna, J. M., & Ventura, S. (2013). Predicting students’ final perfor-
mance from participation in on-line discussion forums. Computers and Education. https​://doi.
org/10.1016/j.compe​du.2013.06.009.
Romero, C., & Ventura, S. (2007). Educational data mining: A survey from 1995 to 2005. Expert Sys-
tems with Applications. https​://doi.org/10.1016/j.eswa.2006.04.005.
Salloum, S. A., Al-Emran, M., Shaalan, K., & Tarhini, A. (2019). Factors affecting the E-learning
acceptance: A case study from UAE. Education and Information Technologies, 24(1), 509–530.
https​://doi.org/10.1007/s1063​9-018-9786-3.
Shahiri, A. M., Husain, W., & Rashid, N. A. (2015). A review on predicting student’s performance using
data mining techniques. Procedia Computer Science. https​://doi.org/10.1016/j.procs​.2015.12.157.

13
 598 A. Abu Saa et al.

Wook, M., Yusof, Z. M., & Nazri, M. Z. A. (2017). Educational data mining acceptance among under-
graduate students. Education and Information Technologies. https​://doi.org/10.1007/s1063​
9-016-9485-x.
Xing, W., Guo, R., Petakovic, E., & Goggins, S. (2015). Participation-based student final perfor-
mance prediction model through interpretable genetic programming: Integrating learning analyt-
ics, educational data mining and theory. Computers in Human Behavior. https​://doi.org/10.1016/j.
chb.2014.09.034.
Yadav, S., Bharadwaj, B., & Pal, S. (2012). Data mining applications: A comparative study for predict-
ing student’s performance. International Journal of Innovative Technology and Creative Engineer-
ing, 1, 13–19.
Yadav, S. K., & Pal, S. (2012). Data mining: A prediction for performance improvement of engineer-
ing students using classification. World of Computer Science and Information Technology Journal
WCSIT. https​://doi.org/10.1142/97898​12771​728_0012.
Yukselturk, E., Ozekes, S., Türel, Y. K., Education, C., Ozekes, S., Türel, Y. K., et al. (2014). Predicting
dropout student: An application of data mining methods in an online education program. European
Journal of Open, Distance and E-Learning. https​://doi.org/10.2478/eurod​l-2014-0008.
Zafra, A., & Ventura, S. (2012). Multi-instance genetic programming for predicting student perfor-
mance in web based educational environments. Applied Soft Computing. https​://doi.org/10.1016/j.
asoc.2012.03.054.
Zhou, Q., Zheng, Y., & Mou, C. (2015). Predicting students’ performance of an offline course from their
online behaviors. In 2015 5th international conference on digital information and communication tech-
nology and its applications, DICTAP 2015. https​://doi.org/10.1109/DICTA​P.2015.71131​73.

Publisher’s Note  Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

13