Case Study of UPNM Students Performance Classification Algorithms

This study applies three classification techniques on educational datasets to predict students’ performance based on coursework assessments. The objective is to improve teaching and learning process by helping lecturers and students identify weak and low performance students.

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/329528133
Case Study of UPNM Students Performance Classification Algorithms
Article in International Journal of Engineering and Technology · December 2018
DOI: 10.14419/ijet.v7i4.31.23382
CITATIONS
0
READS
139
3 authors, including:
Some of the authors of this publication are also working on these related projects:
Situational awareness for malaysian military observersView project
Keyword Patterns Analysis on Military Knowledge using Data Mining Technique for Military PersonnelView project
Nur Diyana Kamarudin
National Defence University of Malaysia
12PUBLICATIONS15CITATIONS
SEE PROFILE
Zuraini Zainol
National Defence University of Malaysia
46PUBLICATIONS106CITATIONS
SEE PROFILE
All content following this page was uploaded by Zuraini Zainol on 10 December 2018.
The user has requested enhancement of the downloaded file.

Copyright © 2018 Authors. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original work is properly cited.
International Journal of Engineering & Technology, 7 (4.31) (2018) 285-289
International Journal of Engineering & Technology
Website: www.sciencepubco.com/index.php/IJET
Research paper
Case Study of UPNM Students Performance Classification
Algorithms
Syarifah B. Rahayu1*, Nur D. Kamarudin2, Zuraini Zainol3
1Cyber Security Centre, National Defence University of Malaysia, Sungai Besi Camp 57000 Kuala Lumpur, Malaysia
2Computer Science Department, Faculty of Science and Defence Technology, National Defence University of Malaysia, Sungai Besi
Camp 57000 Kuala Lumpur, Malaysia
*Corresponding author E-mail: [e-mail]
Abstract
Most students have a problem to keep track on their learning performance. Some lecturers with high teaching hours and burden of ad-
ministration jobs may have difficulty to identify weak and low performance students. In this study, three classification techniques are
applied on educational datasets to predict the students’ performance based on coursework assessments. Thus, this prediction results may
help lecturers and students to improve their teaching and learning process. The objective of study is to predict students’ performance
based on coursework assessments using classification algorithms. The selected classification algorithms applied in this study such as J48
Decision Tree, Naïve Bayes and kNN. WEKA is used as an experimental tool. The selected algorithms are applied on a data of student
database of Data Mining subject. Findings shows Naïve Bayes outperforms other classification algorithms with above 80% prediction
rate. Thus, the students’ performance for Data Mining Subject is improved. As a conclusion, the classification algorithms can predict
students’ performance on a particular subject based on coursework assessments.
Keywords: Prediction; Comparative Analysis; Educational Data Mining
1. Introduction
Educational Data Mining (EDM) researches use data mining tools
to process large quantities of data to discover meaningful patterns
in order to predict students’ performances to enhance teaching and
learning outcomes. These researches can also be used as a plat-
form to alert student on the risk of failure and to provide recom-
mendations for student improvement in their learning process.
One of the criteria for a high quality university is based on its
excellent record of academic achievement [1]. Therefore, student
performance is a crucial part in higher learning institution. A stu-
dent performance is often measured based on the subject work
assessments and final exam. The proposed methodology is to ana-
lyze students’ performance of a particular subject. The findings
are used for predicting their performance before they are taking a
final exam. Thus, it will assist the lecturers or educators to identify
students who need supports to perform well in the final exam.
Besides, students can improve their learning process in order to
pass the subject [2]. The objective of this study is to predict the
students’ performance based on Malaysia Grading System. These
performances are predicted using three different classification
algorithms, for example, J48 Decision Tree, Naïve Bayes and
kNN.
The rest of this paper is organized as follows. Section 2 presents
the background and related work to this study. In Section 3, we
described the framework of our proposed research. Section 4 dis-
cusses the experiment and results. Finally, we conclude this paper
with future work in section 5.
2. Background and Related Works
In this section, some related topics on data mining, knowledge
discovery in databases, classification algorithms and reviews on
related work are discussed.
2.1. Data Mining and Knowledge Discovery in Database
Data Mining (DM) and Knowledge Discovery in Databases (KDD)
are two terms that are often used interchangeably. KDD can be
defined as a process of finding useful information and patterns in
data [3]. In KDD, DM is placed in the fourth steps of the KDD
process. Technically, the KDD process consists of five main steps
such as selection, pre-processing, transformation, data mining, and
interpretation or evaluation (see Fig 1).
According to [3], DM is often applied to extract hidden informa-
tion and useful patterns using algorithms from massive amounts of
data which is derived by the KDD process. Such valuable infor-
mation and patterns may assist the top level managers in decision
making. DM has been applied in various application areas such as
market based analysis, healthcare [5], smart homes [6], business,
text documents [7-10], environmental studies [11, 12], flood de-
tection [13], crime investigation, fraud detection, geology, food
microbiology, astronomy, etc. Researchers [14] summarized some
common data mining tasks and techniques (see Table 1). These
tasks and techniques can be applied individually or they can be
combined together to perform more sophisticated processes.

286 International Journal of Engineering & Technology
Fig. 1: Knowledge Discovery in Databases (KDD) process adopted from [4]
2.2. Classification Algorithms
Classification is a supervised learning where the classes are often
determined before data can be mined [3]. Technically, classifica-
tion will assign the data into several predefined classes. Classifica-
tion technique is often applied for predicting or describing dataset
or nominal categories. Each classification technique (see Table 1)
will apply a learning algorithm to identify a model which is best
fitted the relationship between the set of attributes and the class
label (predefined class) of the input data. The model that has been
produced by a learning algorithm should be able to fit the input
data and predict the class label of the records correctly [15].
Table 1: Data Mining Tasks and Techniques Adopted From [14]
DM Tasks DM Techniques
Classification Decision Tree Induction, Bayesian Classifica-
tion, Fuzzy Logic, Support Vector Machines
(SVM), k-Nearest Neighbors (K-NN), Rough
Set Approach, Genetic Algorithm (GA), etc.
Clustering Partitioning Methods, Hierarchical Methods,
Density-based Methods, Grid-based Methods,
etc.
Association Rules Frequent Item set Mining Methods (e.g.,
Apriori, FP-Growth)
Some examples of classification technique are detecting spam
email messages based on the message header and content, catego-
rizing cells as malignant or benign based on the result of MRI,
identifying credit risks based on bank loan, predicting students’
performance, etc.
In [16-19], the Decision Tree (J48), Bayesian Classifier and k-
Nearest Neighbor (kNN) classifiers have been implemented to
evaluate students’ performances based on several observational
attributes such as accumulated exam grades, percentages or clas-
ses (i.e distinction, fail etc). Based on comparative analysis of
classifier in [16], the Bayesian classifier outperformed the deci-
sion tree and kNN classifier on predicting students’ performances
via average True Positive (TP) rate. However, in analyzing the TP
rate for each classes (Distinction, First, Second, Third and Fail); it
has been observed that, the prediction rates are not uniform among
classes. Hence, the gap of prediction rate among classes is varied
almost 90% in some cases. This might be due to the insufficient
data of certain classes especially in distinction and fail classes.
To discover the optimal classification model for decision tree,
research in [20] did the comparison of different algorithms com-
prises of J48, ID3, C4.5, REPTree, Random Tree and Random
Forest. Out of six decision tree algorithms, the highest percentage
is achieved using the model relying on the algorithm J48. Based
on 161 questionnaires, two researchers from University of Basrah
have analyzed and assisted academic achievers in higher education
using Bayesian Classification Method [21]. For attribute selection,
questions with high correlation averages have been adopted to
enhance the accuracy of classification.
Recent work has been done to demonstrate the efficiency of Semi-
Supervised Learning (SSL) methods for the performance predic-
tion of high school students using their final examination assess-
ment percentage [22]. In this work, various SSL algorithms such
as Self-training, Co-training, Democratic Co-learning, Tri-
training, De-Tri training and RASCO are implemented in KEEL
Software tool. In addition, Friedman Aligned Ranks nonparamet-
ric test is used to measure the performances of these algorithms.
Moreover, in second phase of experiments, the performance of
SSL classifiers have been compared with supervised method, Na-
ïve Bayes. From the observation, it can be concluded that SSL
algorithm are comparatively better than the respective supervised
algorithm, Naïve Bayes based on both measurement; the accuracy
and Friedman Aligned Rank.
A comprehensive survey is then carried out by the Indian Re-
searchers to discuss about the current approaches and potential
areas in EDM [23]. This paper reported the details of researches
done in the area of education in tabular form describing methodol-
ogies and findings of each research and identifies potential re-
search areas for future scope. Similar research is conducted by the
researcher in [24] where the new potential domains of EDM have
been proposed. According to this paper, EDM data is not limited
to predict the student’s performance but can also be utilized in
other domains of education sector (i.e. optimization of resources
or human resource purposes).
In comparison of correlation among pre and post enrollment fac-
tors and employability using data mining tools, many of today’s
graduates are lacking interpersonal communication skills, creative
and critical thinking, problem solving, analytical skills, and team
work [25]. It has been concluded that cognitive factors such as set
of behaviors, skills and attitudes play a significant role in predic-
tion of student’s marketability after graduation. Another work
presented by Research Group for Work, Organizational, and Per-
sonnel Psychology, Department of Psychology, KU Leuven, Leu-
ven, Belgium stated that employability is in strong correlation
with competences and dispositions [26].
Motivated by the previous researches, this research attempts to
evaluate the performances of several students by measuring their
subject work assessment percentages (Quizzes, Tutorials and Test)
via predetermined classes endorsed by the university and Malaysi-
an Grading System to predict their performance in final exam. To
the best of our knowledge, this is the only research paper that
discusses the student’s performance prediction in Malaysia based
on Malaysian Grading System apart from using the students’
CGPA. We proposed this research as a preliminary assessment
tool where we narrower the scope of research to cater early recog-
nition of student who needs help in certain subject not the whole
performances of student from his or her CGPA. In addition, by
analyzing the distributed rank or weightage on each assessment
using decision tree, this research will offer guidance to a lecturer
on improving the teaching plan based on the learning outcomes in
certain assessment as well as to identify weak students to improve
the students’ learning process prior to the final exam. Future con-
tribution will be the automatic application on a platform that is
able to read, analyze and predict the outcome of student’s progress
based on certain assessments in difficult or challenging university
subjects for intelligent tutoring or lecturing applications.
3. Research Framework
Figure 2 illustrates the proposed framework for predicting stu-
dents’ performance in Data Mining subject. The first stage is data
collection. The data about students related to a particular subject

International Journal of Engineering & Technology 287
Fig. 2: A proposed framework of classification model to predict students’ performance on a particular subject
are collected. In this study, the data set is obtained from last se-
mester records of students who are registering for Data Mining
subject. These students are majoring in Computer Science in Sci-
ence Computer Department, National Defence University of Ma-
laysia. For the study case, a sample of 71 students is selected in
this experiment. This study will be used as starting point for a
deep machine learning in the future. Thus, we are focusing on the
smaller dataset before gathering large number of dataset. The Data
Mining subject is consists of 60 marks coursework assessments
and 40 marks final exam. The coursework assessments of this
subject are quizzes, tutorials and test. Students are required to
obtain at least 40 marks to pass the subject.
During the pre-processing stage, the dataset is prepared before
applying the classification algorithms. Then, data attributes are
identified and selected. Table 2 below is the students’ related
attributes. There are 3 quizzes, 3 tutorials and a test. The accumu-
lated values of these attributes are equal to 60 marks. Thus the
calculation will be the values of attributes, t is divided with the
total values of attributes, m and multiplied with marks of assess-
ment coursework, n.
Table 2: Students Related Attributes
Attributes Type Values Grade
Quiz 1 Real [1,10] [A+, F]
Quiz 2 Real [1,10] [A+, F]
Quiz 3 Real [1,10] [A+, F]
Tutorial 1 Real [1,5] [A+, F]
Tutorial 2 Real [1,5] [A+, F]
Tutorial 3 Real [1,5] [A+, F]
Test Real [1,50] [A+, F]
Each coursework assessment including the final exam is graded
based on Malaysia Grading System for university level as tabulat-
ed in Table 3.
Table 3: Malaysia Grading System for University Level
Grade Scale Grade Description
A+ 90.00 – 100.00 Exceptional
A to B+ 76.00 – 89.99 Excellent
B to C+ 65.00 – 75.99 Good
C to D 40.00 – 64.99 Average
F 0.00 – 39.99 Fail
The next stage is to apply classification algorithms on the data set.
This study is using WEKA (Waikato Environment for Knowledge
Analysis) version 3.8.2 as an experimental tool. This tool is devel-
oped at University of Waikato, New Zealand, and known as a
prominent open source data mining tools comprises of several
machine learning classifiers. It has been widely used among re-
searchers and data scientists for data pre-processing, classification,
regression, clustering, association rules, and visualization. Three
algorithms have been selected in this stage. The selected classifi-
cation algorithms are J48 Decision Tree, Naïve Bayes and kNN.
The data set is divided into 10 equally sized folds using the 10-
fold cross validation procedure provided by WEKA (fig 3).
4. Results and Discussion
The objective of this paper is to predict students’ performance in
Data Mining subject. Three classification algorithms are selected
to perform the prediction model.
Test
Data
Training Data
Iteration 1 Test
Data
Iteration 2 Test
Data
Iteration 3 Test
Data
.
.
.
.
.
.
.
.
.
Iteration 10 Test
Data
All Data
Fig. 3: 10-Fold cross validation
4.1. Results of Decision Tree
J48 classification algorithm produces a decision tree (fig 4) with
the size of 15 nodes and 8 leaves. The correctly classified instanc-
es are 56 with 78.8732% for the 10-fold cross-validation testing.
Below is the IF-THEN rules based on the produced decision tree:
IF Test <= 23.7 AND Tutorial 3 <= 4 THEN Class = “Good”
IF Test <= 23.7 AND Tutorial 3 > 4 AND Quiz 1 <= 1.5 THEN
Class = “Good”
IF Test <= 23.7 AND Tutorial 3 > 4 AND Quiz 1 > 1.5 AND Test
<= 15.9 AND Quiz 3 <= 4.5 THEN Class = “Good”
IF Test <=23.7 AND Tutorial 3 > 4 AND Quiz 1 > 1.5 AND Test
<= 15.9 AND Quiz 3 > 4.5 AND Quiz 1 <=3.5 THEN Class =
“Good”
IF Test <=23.7 AND Tutorial 3 > 4 AND Quiz 1 > 1.5 AND Test
<= 15.9 AND Quiz 3 > 4.5 AND Quiz 1 > 3.5 THEN Class =
“Excellent”
IF Test <=23.7 AND Tutorial 3 > 4 AND Quiz 1 > 1.5 AND
Test> 15.9 THEN Class = “Excellent”
IF Test > 23.7 AND Quiz 2 <= 4.5 THEN Class = “Excellent”
IF Test > 23.7 AND Quiz 2 > 4.5 THEN Class = “Exceptional”
Based on the IF-ELSE rules, students who perform better in the
two attributes such as Test, and Quiz 2 are likely to pass the sub-
ject with flying color. Other attributes may cause them to be at
risk. Examples of students data based on J48 Decision Tree (table
4).
Table 4: Example of Student Data based on J48 Decision Tree
Test Quiz 2 Tutorial 3 Class
Student C 24.9 5.0 5.0 Exceptional
Student B 22.2 5.0 5.0 Excellent
Student B 14.1 5.0 4.0 Good
Table 5 below shows the classification results for the decision tree
algorithm. The highest True Positive (TP) Rate is recorded in
Excellent class (91.7%) while the lowest TP rate is recorded in
Good class (42.9%). This maybe due to high number of students
who fall in the Excellent grade for coursework assessments. How-
ever, the weighted average of TP Rate for all observational classes
is 78.9%. In addition, the Precision is high for two classes: Good
(85.7%) and Excellent (80%), and low for Exceptional (66.7%).
The weighted average of Precision is 79.4%.
Data Collection Pre-processing EvaluationData Mining
Raw Data Dataset Attribute
Selection
Classification
Algorithms Findings

288 International Journal of Engineering & Technology
Fig. 4: J48 tree visualization
Table 5: Classification Results for the Decision Tree Algorithm (J48)
Class TP
Rate
FP
Rate
Precision Recall F-Measure
Excellent 0.917 0.478 0.800 0.917 0.854
Exceptional 0.667 0.048 0.667 0.667 0.667
Good 0.429 0.018 0.857 0.429 0.571
Weighted
Avg.
0.789 0.333 0.794 0.789 0.775
4.2. Results of Naïve Bayes
On the other hand, Naïve Bayes classifier has correctly classified
57 instances with 80.2817 % TP rate by using 10-fold cross-
validation testing.
Table 6 below shows the classification results for Naïve Bayes
algorithm. The True Positive (TP) Rate is the highest for the Ex-
ceptional class (100%). However, the TP Rate is lower in two
classes; Excellent (79.2%) and Good (71.4%).The weighted aver-
age of TP Rate is recorded as 80.3%. The Precision is high for
Excellent class (90.5%), and low for: Good (66.7%) and Excep-
tional (64.3%) classes. The weighted average of Precision is
82.5%.
Table 6: Classification Results for the Naïve Bayes Algorithm
Class TP
Rate
FP
Rate
Precision Recall F-
Measure
Excellent 0.792 0.174 0.905 0.792 0.844
Exceptional 1.000 0.081 0.643 1.000 0.783
Good 0.714 0.088 0.667 0.714 0.690
Weighted
Avg.
0.803 0.145 0.825 0.803 0.806
4.3. Results of kNN
kNN classifier correctly classified 53 instances with 74.6479% TP
rate by using 10-fold cross-validation testing.
Table 7 below shows the classification results for k-NN algorithm.
The True Positive (TP) Rate is the highest for Excellent class
(87.5%). The TP Rate of Exceptional Class is recorded as 77.8%
and the lowest TP rate is in Good class (28.6%). However, the
weighted average of TP Rate is 74.6%. Two classes have highest
Precision: Excellent (77.8%) and Exceptional (77.8%). The class
with low Precision is Good (50%). However the weighted average
of Precision is 72.3%.
Table 7: Classification Results for the kNN Algorithm
Class TP Rate FP Rate Precision Recall F-Measure
Excellent 0.875 0.522 0.778 0.875 0.824
Exceptional 0.778 0.032 0.778 0.778 0.778
Good 0.286 0.070 0.500 0.286 0.364
Weighted Avg. 0.746 0.371 0.723 0.746 0.727
4.4. Comparison Results for the kNN algorithm
By comparative analysis, the results of the classification algo-
rithms reveal that the Naïve Bayes performs better than the other
algorithms. Furthermore, Naïve Bayes produces the highest classi-
fication accuracy for Exceptional class. Naïve Bayes produces
accuracy above 80% for all observational classes. The worst clas-
sification accuracy is recorded in J48 and kNN for Good class,
which is below 50%. This is due to the number of tested data ap-
plied on each class. It has been observed that, higher number of
sample data resulted in better classification accuracy. The
weighted average for overall accuracy of all tested classifiers is
well above 70%. Table 8 shows the performance of each classifi-
cation algorithms.
Table 8: Classification Algorithm Performance
J48 Decision Tree Naïve Bayes k-NN
TP Rate 0.789 0.803 0.746
Precision 0.794 0.825 0.723
This finding assists lecturers to identify weak students and help
them to improve their marks. Weak students may also identify
their weaknesses and change their study learning. Table 9 shows
the comparison of results before and after final exam. Number of
students who had grade C to D+ before the final exam was re-
duced from 22 students to 4 students only. Students who had
grade A to B+ and B to C+ increased from 21 students to 35 stu-
dents and 25 students to 30 students, respectively. Overall the
results show students are getting better results after the final exam.
Table 9: Students’ Study Performance
Grade Coursework Final Exam
A+ 3* 2*
A to B+ 21* 35*
B to C+ 25* 30*
C to D 22* 4*
F - -

International Journal of Engineering & Technology 289
5. Conclusion
In this paper, three classification algorithms have been imple-
mented on students’ subject databases to identify their perform-
ance in Data Mining subject based on coursework assessments.
The findings show Naïve Bayes produces the highest classification
accuracy for Exceptional class. Overall weighted average for all
algorithms is above 70%.
This study helps both students and lecturers to enhance their learn-
ing and teaching methods. Students who have low marks in their
subject perform better during the final exam since remedial and
necessary action are taken to enhance their learning process. At
the same time, lecturers give better supports on certain assessment
to improve their students’ performance. As a result, the overall
students’ performance is improved. This prediction system may
embed into e-learning or tutoring system, where students can
measures their study performance.
Future work is to identify other attributes that may contribute stu-
dents’ performance. Furthermore, Next study it will involve larger
educational dataset and different types of classification algorithms
such as Support Vector Machine (SVM), REPTree, RandomTree
and LMT to predict the students’ performance during their under-
graduate studies.
References
[1] Shahiri AM & Husain W, "A review on predicting student's per-
formance using data mining techniques," Procedia Computer Sci-
ence, vol. 72, (2015), pp. 414-422.
[2] Dunham MH, Data mining: Introductory and Advanced Topics,
Pearson Education India, (2006).
[3] Fayyad U, Piatetsky-Shapiro G & Smyth P, "The KDD process for
extracting useful knowledge from volumes of data," Communica-
tions of the ACM, Vol. 39, No. 11, (1996), pp. 27-34.
[4] Azam M, Loo J, Naeem U, Khan S, Lasebae A, & Gemikonakli O,
"A framework to recognise daily life activities with wireless prox-
imity and object usage data," 23rd International Symposium on
Personal, Indoor and Mobile Radio Communication, Sydney, Aus-
tralia, (2012), pp. 590-595.
[5] Jothi N, Abdul RN & Husain W, Data Mining in Healthcare – A
Review, Procedia Computer Science,Vol. 72, (2015), pp. 306-313.
[6] Rashid MM, Gondal I, & Kamruzzaman J, "Mining associated pat-
terns from wireless sensor networks," IEEE Transactions on Com-
puters, Vol. 64, No. 7, (2015), pp. 1998-2011.
[7] Zainol Z, Marzukhi S, Nohuddin PNE, Noormaanshah WMU &
Zakaria O, "Document Clustering in Military Explicit Knowledge:
A Study on Peacekeeping Documents," in Advances in Visual In-
formatics: 5th International Visual Informatics Conference, IVIC
2017, Bangi, Malaysia, November 28–30, 2017, H. Badioze Zaman
et al., Eds. Cham: Springer International Publishing, (2017), pp.
175-184.
[8] Zainol Z, Nohuddin PNE, Mohd TAT & Zakaria O, “Text analytics
of unstructured textual data: a study on military peacekeeping doc-
ument using R text mining package”. Proceeding of the 6th Inter-
national Conference on Computing and Informatics (ICOCI17),
(2017), pp. 1–7.
[9] Nohuddin PNE, Zainol Z, Chao FC, James MT & Nordin A, "Key-
word based Clustering Technique for Collections of Hadith Chap-
ters," International Journal on Islamic Applications in Computer
Science And Technologies – IJASAT, Vol. 4, No. 3, (2015), pp. 11-
18.
[10] Zainol Z, Nohuddin PNE, Jaymes MTH & Marzukhi S, "Discover-
ing “interesting” Keyword Patterns in Hadith Chapter Documents,"
in Information and Communication Technology (ICICTM16), In-
ternational Conference on, Kuala Lumpur, Malaysia, (2016), pp.
104-108.
[11] Rashid RAA, Nohuddin PNE & Zainol Z, "Association Rule Min-
ing Using Time Series Data for Malaysia Climate Variability Pre-
diction," in Advances in Visual Informatics: 5th International Visu-
al Informatics Conference, IVIC 2017, Bangi, Malaysia, November
28–30, 2017, Proceedings, H. Badioze Zaman et al., Eds. Cham:
Springer International Publishing, (2017), pp. 120-130.
[12] Nohuddin PNE, Zainol Z & Ubaidah MAB, "Trend Cluster Analy-
sis of Wave Data for Renewable Energy," Advanced Science Let-
ters, Vol. 24, No. 2, (2018), pp. 951-955.
[13] Harun NA, Makhtar M, Aziz AA, Zakaria ZA, Abdullah FS &
Jusoh JA, "The Application of Apriori Algorithm in Predicting
Flood Areas," International Journal on Advanced Science, Engi-
neering and Information Technology, Vol. 7, No. 3, (2017).
[14] Nohuddin PNE, Z. Zainol, Lee ASH, Nordin AI & Yusoff Z, "A
case study in knowledge acquisition for logistic cargo distribution
data mining framework," International Journal of Advanced and
Applied Sciences, Vol. 5, No. 1, (2017), pp. 8-14.
[15] Tan PN, Introduction to data mining, Pearson Education India,
(2006).
[16] Anuradha C & Velmurugan T, “A Comparative Analysis on the
Evaluation of Classification Algorithms in the Prediction of Stu-
dents Performance,” Indian Journal of Science and Technology,
Vol. 8, No. 15, (2015).
[17] Goga M, Kuyoro S & Goga N, "A Recommender for Improving the
Student Academic Performance," Procedia - Social and Behavioral
Sciences, Vol. 180, (2015), pp. 1481-1488.
[18] Kabakchieva D, "Predicting Student Performance by Using Data
Mining Methods for Classification," in Cybernetics and Infor-
mation Technologie, Vol. 13(2013), pp. 61.
[19] Hussain S, Abdulaziz DN, Ba-Alwib FM & Najoua R, “Educational
Data Mining and Analysis of Students' Academic Performance Us-
ing WEKA,” Indonesian Journal of Electrical Engineering and
Computer Science, Vol. 9, No. 2, (2018) pp. 447-459.
[20] Mesaric J & Šebalj D, “Decision trees for predicting the academic
success of students,” Croatian Operational Research Review, Vol.
9, No. 2, (2016), pp. 367-388.
[21] Khalaf A, Humadi AM, Akeel W & Hashim AS, “Students’ Suc-
cess Prediction based on Bayes Algorithms,” SSRN elibrary, Else-
vier, (2017), pp. 6-12.
[22] Kostopoulos G, Livieris I, Kotsiantis S & Tampakas V, “Enhancing
high school students’ performance based on semi-supervised meth-
ods,” in 8th International Conference on Information, Intelligence,
Systems & Applications (IISA), (2017), pp. 1-6.
[23] Thakar P, “Performance Analysis and Prediction in Educational
Data Mining: A Research Travelogue,” International Journal of
Computer Applications, Vol. 110, No. 15, (2015), pp. 60-68.
[24] Kumar DV & Anupama C, “An Empirical Study of the Applica-
tions of Data Mining Techniques in Higher Education,” Interna-
tional Journal of Advanced Computer Science and Applica-
tions, Vol.3, No. 2, (2011), pp. 81-84.
[25] Bakar NAA, Mustapha A & Nasir KM, "Clustering analysis for
empowering skills in graduate employability model," Australian
Journal of Basic and Applied Sciences, Vol. 7, No. 14, (2013), pp.
21-24.
[26] Dorien V, Nele DC, Ellen P & Hans DW, "Defining perceived em-
ployability: a psychological approach," Personnel Review, Vol. 43,
No. 4, (2014), pp. 592-605.
View publication statsView publication stats