Using Naive Bayesian Classifier for Predicting Performance of a Student

International Journal of Trend in Scientific Research and Development (IJTSRD)
Volume 3 Issue 5, August 2019 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470
@ IJTSRD | Unique Paper ID – IJTSRD26463 | Volume – 3 | Issue – 5 | July - August 2019 Page 1387
Using Naive Bayesian Classifier for
Predicting Performance of a Student
Khin Khin Lay, Aung Cho
University of Computer Studies, Maubin, Myanmar
How to cite this paper: Khin Khin Lay |
Aung Cho "Using NaiveBayesianClassifier
for Predicting Performance of a Student"
Published in
International
Journal of Trend in
Scientific Research
and Development
(ijtsrd), ISSN:2456-
6470, Volume-3 |
Issue-5, August
2019, pp.1387-1391,
https://doi.org/10.31142/ijtsrd26463
Copyright © 2019 by author(s) and
International Journal ofTrend inScientific
Research and Development Journal. This
is an Open Access article distributed
under the terms of
the Creative
Commons Attribution
License (CC BY 4.0)
(http://creativecommons.org/licenses/by
/4.0)
ABSTRACT
Data mining techniques play an important role in data analysis. For the
construction of a classification model which could predict performance of
students, particularly for engineering branches, a decision tree algorithm
associated with the data mining techniques have been used in the research. A
number of factors may affect the performance of students. Data mining
technology which can relate to this student grade well and we also used
classification algorithms prediction. We proposed student data classification
using Naive Bayesian Classifier.
KEYWORDS: Classification, NaiveBayesian, DataMining, PredictingPerformance
I. INTRODUCTION
Data Mining Advancement in technology has brought great growth in the
volume of the data available on the internet, digital libraries, news sources and
company-wide intranets. This makes a huge number of databases and
information repositories available, but it is impossible to manually organize,
analyze and retrieve information from this data.
This generates an eminent need of methods that can help users to efficiently
navigate, summarize, and organize the data so that it can further be used for
applications ranging from market analysis, fraud detection and customer
retention etc. Therefore, the techniques that perform data analysis and may
uncover important data patterns are needed. One of these techniques is known
as data mining.
There are increasing research interests in using data mining
in education. This new emerging field, called Educational
Data Mining, concerns with developing methods that
discover knowledge from data originating from educational
environments [4]. Educational Data Mining uses many
techniques such as Decision Trees, Neural Networks, Naïve
Bays, K- Nearest neighbor, and many others.
The main aim of this paper is to use data mining
methodologies to study students’ performance in the
courses. Data mining provides many tasksthatcouldbeused
to study the student performance. Here the classification
tasks is used to evaluate student’s performance and as there
are many approaches that are used for data classification,
the decision tree and Naive Bays method are used [4].
Decision trees can easily be converted to classification rules
Decision tree algorithms, such as ID3 (Iterative
Dichotomies), C4.5, and CART (ClassificationandRegression
Trees) [3]. Naïve Bayesian classifiers assume that the effect
of an attribute value on a given class is independent of the
values of the other attribute [3].This paper explores the
accuracy of Decision tree and Naive Bays techniques for
predicting student performance.
II. RELATED WORK
In order to predict the performance of students thesearcher
took into consideration the work of other14ADecision Tree
Approach forPredicting Students Academic Performance
researchers that are in the samedirection.Otherresearchers
have looked at the work of predicting students’performance
by applying many approaches and coming up with diverse
results.
Classification is a classic data mining technique based on
machine learning. Basically classification is used to classify
Each item in a set of data into one of predefinedset ofclasses
or groups. Classification method makes use of mathematical
Techniques such as decision trees, linear programming,
neural network and statistics. In classification, once the
software is made that can learn how to classify the data
items into groups.
Three supervised data mining algorithms, i.e. Bayesian,
Decision trees and Neural Networks which were applied by
[1] on the preoperative assessmentdata topredictsuccess in
a course (to produce result as either passed or failed) and
the performance of the learning methods were evaluated
based on their predictive accuracy, ease of learninganduser
friendly characteristics. The researchers observed that that
this methodology can be used to help students and teachers
to improve student’s performance; reduce failing ratio by
taking appropriate steps at right time to improve the quality
of learning.
The authors in [9] analyzed student’s performance data
using classification algorithmnamedID3topredictstudent’s
marks at the end of the semester. This was applied for
master of computer applications course from 2007 to 2010
in VBS Purvanchal University, Jaunpur. Their studyaimedto
help students and teachers find ways to improve students’
IJTSRD26463

International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
performance. Data was collected from 50 students, and then
a set of rules was extracted for their analysis. Another study
that focused on the behavior to improve students'
performance using data mining techniques is illustrated in
[10]. The data consisted of 151 instances from a data base
management system course held at the Islamic University of
Gaza. The data was collected from personal records and
academic records of students. The author performed the
data mining techniques, namely: association rules,
classification, and clustering and outlier detection. The
results revealed useful information from association rules
and classification models.
III. OUR PROPOSED METHOD
a. Naïve Bayesian Classifier
Student performance is predicted using a data mining
technique called classification rules. The naïve Bayes
classification algorithm is used by the administrator to
predict student performance based on performance detail.
The algorithm is a simple probabilistic classifier that
calculates a set of probabilities by counting the frequency
and combinations of values in a given dataset.Thealgorithm
uses the Bayes theorem and assumes that all attributes are
independent given the value of the class variable. This
conditional independence assumption rarely holds true in
real-world applications, hence the characterizationas naïve.
However, the algorithm tends to perform well and learn
rapidly in various supervised classification problems [11].
The naïve Bayesian classifier works as follows:
1. We let T be a training set of samples, each with their
class labels. k classes, C1,C2, . . . ,Ck exist. Each sample is
represented by an n-dimensional vector, X = {x1, x2, . . . , xn}
that depicts n measured values of the n attributes, A1,A2, . .. ,
An, respectively.
2. Given a sample X, the classifier will predict that X
belongs to the class having the highest a posteriori
probability,
Conditioned on X. X is predicted to belong to the class Ci if
and only if
P(Ci|X) > P(Cj |X) for 1<= j <= m, j ≠i.
Thus, we find the class that maximizes P(Ci|X). The class Ci
for which P(Ci|X) is maximized is called the maximum
posteriori hypothesis. By Bayes’ theorem,
P(Ci|X) = P(X|Ci) P(Ci) P(X)
3. Given that P(X) is the same for all classes, only
P(X|Ci)P(Ci) needs to be maximized. If the class a priori
probabilities, P(Ci), are not known, then we assume that the
classes are equally likely, that is, P(C1) = P(C2) = . . . = P(Ck),
and we would therefore maximize P(X|Ci). Otherwise, we
maximize P(X|Ci)P(Ci). The class a priori probabilities may
be estimated by P(Ci) = freq(Ci, T)/|T|.
4. Given data sets with many attributes, computing P(X|Ci)
would be computationallyexpensive.Toreducecomputation
in evaluating P(X|Ci)P(Ci), the naïve assumption of class
conditional independence is made. This assumption
presumes that the values of the attributes are conditionally
independent of one another, given the class label of the
sample.
The probabilities P(x1|Ci), P(x2|Ci), . . . , P(xn|Ci) can easily
be estimated from the training set. Here, xk refers to the
value of attribute Ak for sample X.
5. To predict the class label of X, P(X|Ci)P(Ci) is evaluatedfor
each class Ci. The classifier predicts that theclass labelofXis
Ci if and only if it is the class that maximizes P(X|Ci)P(Ci)
[12].
IV. Training Dataset
The first step in this paper is to collect data. Itis important to
select the most suitable attributes which influence the
student performance. We have training set of 30 under
graduate students. We were provided witha trainingdataset
consisting of information about students admitted to the
first year in Table I.
TABLE.I Training dataset
Sr. no. Roll no. Attend-ance Apti- tude Assign-ment Test Present-ation Grade
1 IT1 Good Avg Yes Pass Good Excellent
3 IT 3 Good Avg Yes Pass Good Excellent
6 IT6 Avg Avg Yes Pass Avg Good
7 IT7 Poor Good Yes Pass Avg Good
8 IT8 Avg Good Yes Pass Avg Good
10 IT10 Poor Poor No Fail Poor Fail
12 IT12 Avg Age Yes Pass Age Excellent
13 IT13 Good Good Yes Pass Good Excellent
20 IT20 Good Poor Yes Pass Good Excellent
P(XCi)≈∏ 𝑃𝑛
𝑘=1 (XkCi) Eq(1.1)

27 IT27 Poor Good No Fail Poor Fail
A. Classification Steps
TABLE.II Frequency Tables
Frequency Table
Grade
Excellent Good Fail
Attendance
Good 20 0 0
Average 1 4 0
Poor 0 1 4
Frequency Table
Grade
Excellent Good Fail
Aptitude
Good 7 4 1
Average 9 1 0
Poor 5 0 3
Frequency Table
Grade
Excellent Good Fail
Assignment
Yes 21 5 0
No 0 0 4
Frequency Table
Grade
Excellent Good Fail
Test
Pass 20 5 0
Fail 0 0 4
Frequency Table
Grade
Excellent Good Fail
Presentation
Good 20 0 0
Average 1 5 0
Poor 0 0 4
TABLE.III Likelihood Tables
Likelihood Table
Grade
Excellent Good Fail
Attendance
Good 20/21 0/5 0/4 20/30
Average 1/21 4/5 0/4 5/30
Poor 0/21 1/5 4/4 5/30
21/30 5/30 4/30
Likelihood Table
Grade
Excellent Good Fail
Aptitude
Good 7/21 4/5 1/4 12/30
Average 9/21 1/5 0/4 10/30
Poor 5/21 0/5 3/4 8/30
21/30 5/30 4/30
Likelihood Table
Grade
Excellent Good Fail
Assignment
Yes 21/21 5/5 0/4 26/30
No 0/21 0/5 4/4 4/30
21/30 5/30 4/30

Likelihood Table
Grade
Excellent Good Fail
Test
Pass 21/21 5/5 0/4 26/30
Fail
0/21
0/5 4/4 4/30
21/30 5/30 4/30
Likelihood Table
Grade
Excellent Good Fail
Presentation
Good 20/21 0/5 0/4 20/30
Avg 1/21 5/5 0/4 6/30
Poor 0/21 0/5 4/4 4/30
21/30 5/30 4/30
We computed all possible individual probabilities conditioned on the target attribute (Grade) in Table IV.
TABLE IV Possible Probabilities
P(Grade=Excellent) 21/30=0.7
P(Atten=GoodGrade=Ex) 0.9523
P(Atten=AvgGrade=Ex) 0.0476
P(Atten=PoorGrade=Ex) 0
P(Aptitu=GoodGrade=Ex) 0.3333
P(Aptitu=AvgGrade=Ex) 0.4286
P(Aptitu=PoorGrade=Ex) 0.2381
P(Assig=YesGrade=Ex) 1
P(Assig=NoGrade=Ex) 0
P(Test=PassGrade=Ex) 1
P(Test=FailGrade=Ex) 0
P(Present=GoodGrade=Ex) 0.9523
P(Present =AvgGrade=Ex) 0.0476
P(Present =PoorGrade=Ex) 0
P(Grade=Good) 5/30=0.1667
P(Atten=GoodGrade=G) 0
P(Atten=AvgGrade=G) 0.8
P(Atten=PoorGrade=G) 0.2
P(Aptitu=GoodGrade=G) 0.8
P(Aptitu=AvgGrade=G) 0.2
P(Aptitu=PoorGrade=G) 0
P(Assig=YesGrade=G) 1
P(Assig=NoGrade=G) 0
P(Test=PassGrade=G) 1
P(Test=FailGrade=G) 0
P(Present=GoodGrade=G) 0
P(Present =AvgGrade=G) 1
P(Present =PoorGrade=G) 0
P(Grade=Fail) 4/30=0.133
P(Atten=GoodGrade=Fail) 0
P(Atten=AvgGrade=Fail) 0
P(Atten=PoorGrade=Fail) 1
P(Aptitu=GoodGrade=Fail) 0.25
P(Aptitu=AvgGrade=Fail) 0
P(Aptitu=PoorGrade=Fail) 0.75
P(Assig=YesGrade=Fail) 0
P(Assig=NoGrade=Fail) 1
P(Test=PassGrade=Fail) 0
P(Test=FailGrade=Fail) 1
P(Present=GoodGrade=Fail) 0
P(Present =AvgGrade=Fail) 0
P(Present =PoorGrade=Fail) 1

b. Testing Data
Sr.no. Roll no. Attend-ance Apti- tude Assign-ment Test Presentation Grade
1 IT1 Good Avg Yes Pass Good ?
2 IT27 Poor Good No Fail Poor ?
Likelihood of “Excellent” on that IT1
= P (Atten=GoodGrade=Ex) * P(Aptitu=AvgGrade=Ex)*
P (Assig=YesGrade=Ex) * P(Test=PassGrade=Ex) *
P (Present=GoodGrade=Ex) * P(Grade=Excellent)
= 0.9523* 0.4286*1*1* 0.9523* 0.7 = 0.27208
Similarly Likelihood of “Good” on that IT1 =0
Similarly Likelihood of “Fail” on that IT1 =0
Similarly Likelihood of “Excellent” on that IT27 =0
Similarly Likelihood of “Good” on that IT27 =0
Likelihood of “Fail” on that IT27
= P (Atten=PoorGrade=Fail) * P(Aptitu=GoodGrade=Fail)
* P (Assig=NoGrade=Fail) * P(Test=FailGrade=Fail)
* P(Present =PoorGrade=Fail * P(Grade=Fail)
= 1 * 0.25 * 1 * 1* 1* 0.133 = 0.03325
Probability of IT1 being Grade = “Excellent” is 0.27208
Which is greater than anotherprobability?Hence,IT1willbe
Grade to Excellent.
Probability of IT27 being Grade =” Fail” is 0.03325which is
greater than another probability. Hence, IT27 will be Grade
to Fail.
V. CONCLUSION
A classification model has been proposed in this study for
predicting student’s grades particularly for IT under
graduate students. In this paper, the classification task is
used on student database to predict the students divisionon
the basis of previous database. As there are many
approaches that are used for data classification, Naive
Bayesian classifiers method is used here. Information’s like
Attendance, aptitude, test, Presentation and Assignment
marks were collected from the student’s previous database,
to predict the performance at the end of the semester. Bayes
classifier algorithms are used in this model and theaccuracy
of prediction is compared to find theoptimal one.Finally,the
naïve Bayes algorithm is selected as the best algorithm for
prediction based on performance detail.
REFERENCES
[1] Osmanbegovic E., Suljic M. “Data mining approach for
predicting student performance” Economic Review-
Journal of Economics and Business. Volume 10(1)
(2012)
[2] Behrouz, M, Karshy, D, Korlemeyer G, Punch, W.
“Predicting student performance: an application of
data Mining methods with the educational web-based
system” Lon-capa. 33rd ASEE/IEEE Frontiers in
Education Conference. Boulder C.O. USA, (2003).
[3] Bekele, R., Menzel, W. “A bayesian approach to predict
performance of a student (BAPPS): A Case with
Ethiopian Students”. Journal of Information Science
(2013).
[4] Han and M. Kamber, “Data Mining: Concepts and
Techniques, Morgan Kaufmann, 2000
[5] [5] Surjeet K, Yadav, Bharadwaj, B. Pal B.” Data Minig
Applications: A comparative Study for Predicting
Student’s performance.” International journal of
innovative technology & creative engineering. Volume
1(12). (2012).
[6] Nnamani, C. N, Dikko, H. G and Kinta, L. M. “Impact of
students‟ financial strength on their academic
performance: Kaduna Polytechnic experience”.African
Research Review 8(1), (2014).
[7] Ogunde A.O., Ajibade D.A. “A data Mining System for
Predicting University Students F=Graduation Grade
Using ID3 Decision Tree approach”, Journal of
Computer Science and Information Technology,
Volume 2(1) (2014).
[8] Undavia, J. N., Dolia, P. M.; Shah, N. P. “Prediction of
Graduate Students for Master Degree based on Their
Past Performance using Decision Tree in Weka
Environment”. International Journal of Computer
Applications; Volume 74 (21), (2013).
https://www.wekatutorial.com/
[9] B. K. Baradwaj and S. Pal, "Mining educational data to
analyze students' performance " International Journal
of Advanced Computer Science and Applications, vol. 2,
2011.
[10] A. El-Halees, "Mining student data to analyze learning
behavior: A case study," presented at the International
Arab Conference of Information Technology
(ACIT2008), Tunisia, 2008.
[11] George Dimitoglou, James A. Adams, and Carol M. Jim,”
Comparison of the C4.5 and a Naïve Bayes Classifierfor
the Prediction of Lung Cancer Survivability”, Journal of
Computing, Volume 4, Issue 8, 2012.
[12] Leung, K. Ming. “Naïve Bayesian classifier.”Polytechnic
University Department of Computer Science/Finance
and Risk Engineering (2007).

Using Naive Bayesian Classifier for Predicting Performance of a Student

More Related Content

What's hot (17)

Similar to Using Naive Bayesian Classifier for Predicting Performance of a Student (20)

More from ijtsrd (20)

Recently uploaded (20)

Using Naive Bayesian Classifier for Predicting Performance of a Student