Prediction of Students Performance using
Educational Data Mining
Ms.Tismy Devasia1 ,Ms.Vinushree T P2, Mr.Vinayak Hegde3
Department of Computer Science
Amrita Vishwa Vidyapeetham University,Mysuru Campus
Mysuru, Karnataka, India
tismydevasia1993@gmail.com, vinushree7792@gmail.com,vinayakhegde92@gmail.com
Abstract—Data mining plays an important role in the
business world and it helps to the educational institution to
predict and make decisions related to the students’
academic status. With a higher education, now a days
dropping out of students’ has been increasing, it affects not
only the students’ career but also on the reputation of the
institute. The existing system is a system which maintains
the student information in the form of numerical values
and it just stores and retrieve the information what it
contains. So the system has no intelligence to analyze the
data. The proposed system is a web based application
which makes use of the Naive Bayesian mining technique
for the extraction of useful information. The experiment is
conducted on 700 students’ with 19 attributes in Amrita
Vishwa Vidyapeetham, Mysuru. Result proves that Naive
Bayesian algorithm provides more accuracy over other
methods like Regression, Decision Tree, Neural networks
etc., for comparison and prediction. The system aims at
increasing the success graph of students using Naive
Bayesian and the system which maintains all student
admission details, course details, subject details, student
marks details, attendance details, etc. It takes student’s
academic history as input and gives students’ upcoming
performances on the basis of semester.
Keywords—Educational Data Mining, Predicting student
performance, Naive Bayes’, Dropout, Classification.
I. INTRODUCTION
Analyzing the huge amount of data to form summarized useful
information is a tedious task for human kind. Data Mining is
the area which analyzes huge repositories of data to extract
necessary or useful information. Computers can process any
kind of data like numbers, texts, images and facts. This task
performs the analysis based on the patterns, association,
relations among all these data so as to get the information.
The prediction with high accuracy in students' performance is
beneficial as it helps in identifying the students with low
academic achievements at the early stage of acdemics. In
universities, student retention is related to academic
performance and enrollment system. The steps to assist the
low academic performers with better education are:
(a) Generation of data source of predictive variables.
(b) Identification of various features or factors which affects
the performance of student’s learning during academic career.
(c) Construction of a prediction model with the help of
classification data mining techniques on the basis of predictive
variables which is readily identified.
(d) Validation of the model which is developed for
Universities with students’ performance.
Data Mining can be put in the educational field to extend our
understanding of learning process by identifying the variables
and evaluating them. Mining in the field of educational
environment is known as the Educational Data Mining.
Students’ attendance in class, hours spent on a daily basis after
college, family income, mother's age and mother's education
of the students are significantly related to student
performance. By means of naive Bayes’ model, it has been
found that the factors like mother’s qualification and income
of the family are highly correlated with the performance of the
student. Data mining techniques economically offer more
customized education, improved system efficiency, and reduce
the education process expenses for universities. This guide us
to extend student retention rate, increase academic
achievements in case of student learning end result. Data
mining prediction technique helps to spot out the most
effective factors which work with student’s test score and then
tune these factors to make better student test performance. It
provides a new way to look into the education system which
was hidden from humankind.
II. RELATED WORK
J K Jothi and K Venkatalakshmi conducted the students’
performance analysis on the graduate students’ data collected
from the Villupuram college of Engineering and Technology.
The data included five year period and applied clustering
methods on the data to overcome the problem of low score of
graduate students, and to raise students academic performance
[1].

Sheik and Gadage have done the analysis related to the student
learning behavior by using different data mining models,
namely classification, clustering, decision tree, sequential
pattern mining and text mining. They used open source tools
such as KNIME (Konstanz Information Miner),
RAPIDMINER, WEKA, CARROT, ORANGE, R-
Programming, and iDA. These tools have different
compatibilities and it provided an insight into the prediction
and evaluvation [2].
Mythili M S and Shanavas A R applied classification
algorithms to analyze and evaluate school students’
performance using weka. They came with various
classification algorithms, namely J48, Random Forest,
Multilayer perception, IBI and decision table with the data
collected from the student management system [3].
Dinesh A and Radhika V targeted on the techniques and
strategies of instructional data processing for data discovery
from the information collected from various universities. This
paper stated that relationship mining was leading between
1995 and 2005 and in 2008 to 2009 it slipped to 5th place.
During the period 2008 to 2015 45% papers are moving to
prediction. The prediction model acts like a warning system to
improve their performance [4].
Osmanbegovic and Suljic conducted a study for investigating
students’ future performance in the end semester results at the
University of Tuzla. They considered 11 factors and used
classification model with highest accuracy for naive Bayes [5].
Suyal and Mohod applied the association and classification
rule to identify the students’ performance. They mainly
focused to find the students who need special attention to
reduce failure rate [6].
Noah, Barida and Egerton conducted a study to evaluate
students’ performance by grouping the grading into various
classes using CGPA. They used different methods like Neural
network, Regression and K-means to identify the weak
performers for the purpose of performance improvement [7].
Baradwaj and pal described data mining techniques that help
in early identification of student dropouts and students who
need special attention. Here they used a decision tree by using
information like attendance, class test, semester and
assignment marks [8].
Jeevalatha, Ananthi, and Saravana Kumar presented a case
study on performance analysis for placement selection for
undergraduate students. They applied decision tree algorithm
by considering the factors like HSC, UG marks and
communication skills [9].
Backer and Yacef conducted a study for identifying the most
appropriate model for EDM. They analyzed data and reached
the conclusion that most of the papers adopt prediction than
relationship mining [10].
ElGamal A F presented a study for predicting student
performance in a programming course. Here the data is
collected from the department of computer science from
Mansoura University and applied extract rules for predicting
students’ performance in programming course [11].
Angeline D M conducted a study on the students’ performance
by using Apriori algorithm that extracts the set of rules
specific to every category and analyze the given knowledge to
classify the scholar based on their involvement in assignment,
internal assessment test, group action etc. It helps to identify
the students’ performance range like average, below average,
and good performance [12].
Bhise, Thorat and Supekar presented a method using K-means
clustering algorithm by describing it step by step. This paper
mainly focused on reducing drop-out-ratio of the students and
improve it by considering the evaluation factors like midterm
and final exam assignment. They considered different
clustering techniques namely hierarchical, partitions, and
categorical [13].
Remesh, Parkavi, and Yasodha conducted a study on the
placement chance prediction by investigating the different
techniques such as Naive Bayes Simple,
MultiLayerPerception, SMO, J48, and REPTree by its
accuracy. From the result they concluded that
MultiLayerPerception technique is more suitable than other
algorithms [14].
Tair M M A and El-Halees presented a case study with a set
of data collected from degree holders of college ‘Science and
Technology, Khanyounis’, during the period of 1993 to 2007.
They used two classification methodologies such as Rule
Induction and Naive Bayesian classifier to forecast the grades
of the students [15].
III. METHODOLOGY
Data mining is the knowledge discovery process from a
huge data volume. The mechanism works in large dataset
where the student performance in the end semester
examination is evaluated.
A. Data preparation
Student related data were collected from the college Amrita
School of Arts and Sciences, Mysuru on the sampling method
of computer science department from the session 2013 to
2016. In this step, data stored in different tables were joined
into a single set.
B. Data selection and transformation
In this step only those fields were selected which were
required for the data mining process. The student register
number, 10th, 12th, degree marks in each semester wise,
assignment, gender, parent’s education, income were taken as
the attribute values for predictions. This is shown in the C. Naive Bayes’ Algorithm
table1.

TABLE I. STUDENT RELATED VARIABLES
SI.No DESCRIPTION
1 Students’ Gender
2 Students’ Category
3 Medium of Teaching
4 Students’ Food Habit
5 Students’ Other Habit
6 Living Location
7 Where do you stay
8 Number of Members in a
Family
9 Students’ Family Status
Family Annual Income status
10
11 Students’ Grade in 10th /
SSLC
12 Students’ Grade in 12th/ PUC
Step 1: Scan the student data set
Step 2: Calculate the probability of each attribute value.
POSSIBLE VALUES [n, n_c, m, p]
Male, Female
Step 3: Apply the formulae
P(attribute value(ai)/subject value (vj))=(n_c + mp)/(n+m)
GM, SC, ST, OBC Where:
KAN, HIN, ENG n = the number of training data item for which v = vj
nc = number of examples for which v = vj and a = ai
Veg, Non-Veg
p = a priori estimate for P(ai,vj)
Drinking, Smoking, NA m = the parallel size of the sample
Village, City Step 4: Multiply the probabilities by p
Step 5: Compare the values and classify the attribute values to
Hostel, Room, PG one of the predefined set of class.
2, 3, >3
Joint, Nuclear IV. EXPERIMENTAL RESULTS
BPL, Poor, Medium, High
<40, 40-59,60-80, >80
<40, 40-59,60-80, >80

13 Students College Type Boys, Girls, Combined

No-Education, Elementary,
14 Father’s qualification Secondary, Graduate,
postgraduate,Doctorate, NA

No-Education, Elementary,
15 Mother’s Qualification Secondary, Graduate,
postgraduate, Doctorate, NA

Farmer, Business, Service,

16 Father’s Occupation Retired,
Not-Applicable
Housewife, Business,
17 Mother’s Occupation Service, Retired,
Not-Applicable

Fig.1. Student Data Set

18 Student Interested In Higher
Yes, No
Education
Above fig.1 represents the students’ data set collected from a
19 Do you use Mobile? If Yes, Yes, No database as well as a survey of approximately 60 students at
Since how many Amrita school of Arts and Sciences, Mysore .
Months/Years
20 Do you use the Internet? If Yes, No
Yes, Since how many
Months/Years
21 Do you use the social Yes, No
network? If Yes, Since how
many Months/Years

22 How many Siblings and their

1, 2, >3
qualification
23 Reading Habit Early morning , Night

24 How many hours do you 1, 2, >3

spend on studies per day
25 Do you use a vehicle Bike, Car, Bicycle, NA

Fig.2. Students 12th mark levels

Fig.2 shows the tenth grade performance of students. By
observation from the graph its clearly evident that among sixty
students half-dozen students area unit belong to the category
below forty, eleven students belong to forty to fifty nine, thirty
one students belong to sixty to eightieth category and twelve
students belong to higher than eighty. The graph displays that
minimum performance of students is thirty seven, maximum is
ninety eight, average students’ performance is seventy nine
and variance of students’ performance is fifteen.
Fig.3. Student Attributes for Prediction
Fig.3 shows the attribute values of a student. When a new
student enters his or her attributes, the values will be
compared with the existing student’s attributes to evaluate the
performance and predict the same. The performance is
predicted using Naive Bayesian algorithm. This helps the
organization to predict and analyze the student performance in
a better manner.
Fig.4. Students performance Prediction Graph
Fig. 4 shows the performace evaluation graph of students in
percentage. It is evident from the graph that most of the
students are good in performance.
VI. CONCLUSION
In this paper, the classification is employed in student
information to predict the students' division on the premise of
previous information. As there are several approaches that
area unit used for knowledge classification, Naive theorem is
employed here. Information like group action, class test,
seminar and assignment marks were collected from the
students’ previous information, to predict the performance at
the top of the semester.
This study can facilitate the students and the lecturers to boost
the students of all category to perform well. This study helps
to spot out those students who require special attention ,
minimize the failure ratio and to take acceptable action for
upcoming semester examination.
Future work includes applying data processing techniques for
Associate in nursing distended knowledge set with additional
typical attributes to urge correct and economical results.
.References
[1] J.K.Jothi and K.Venkatalakshmi, “Intellectual performance analysis of
students by using data mining techniques”, International Journal of
Innovative Research in Science, Engineering and Technology, vol 3,
Special iss 3, March 2014.
[2] Nikitaben Shelke and Shriniwas Gadage, “A survey of data mining
approaches in performance analysis and evaluation”, International
Journal of Advanced Research in Computer Science and Software
Engineering , vol 5, iss 4, 2015K.
[3] M.S. Mythili1 and A.R.Mohamed Shanavas , “An analysis of students’
Performance using classification algorithms ”, IOSR-JCE, Volume 16,
iss1, Jan. 2014.
[4] A.Dinesh Kumar and V.Radhika, “A survey on predicting student
performance”, International Journal of Computer Science and
Information Technologies, Vol. 5, 2014.
[5] E. Osmanbegović and M. Suljić, '”ata mining approach for predicting
students performance”, Economic Review, vol 10, iss 1, 2012.
[6] Sayali Rajesh Suyal and Mohini Mukund Mohod, “Quality
improvisation of student performance using data mining techniques”,
International Journal of Scientific and Research Publications, vol 4,iss 4,
April 2014.
[7] OTOBO Firstman Noah, BAAH Barida and Taylor Onate Egerton,
“Evaluation of student performance using data mining over a given data
space”, International Journal of Recent Technology and Engineering
(IJRTE) ISSN: 2277-3878, Volume-2, iss 4, September 2013.
[8] Brijesh Kumar Baradwaj and Saurabh Pal, “Mining educational data to
analyze Ssudents’ performance”, (IJACSA) International Journal of
Advanced Computer Science and Applications, Vol. 2, No. 6, 2011.
[9] T.Jeevalatha, N. Ananthi and D.Saravana Kumar, “Performance analysis
of undergraduate students placement selection using Decision Tree
Algorithms”, International Journal of Computer Applications (0975-
8887), vol 108, December 2012.
[10] Ryan J.D.B.Baker and Kalina Yacef , “The state of educational data
mining in 2009: A review and future revisions”, Journal of Educational
Data Mining , Vol.1,No.1, February 2009.
[11] A.F . ElGamal, “An educational data mining model for predicting
student performance in programming course”, International Journal of
Computer Applications(0975-8887), Vol.70, No.17, May 2013.
[12] D.Magdalene Delighta Angeline, “Association rule generation for
student performance analysis using Apriori Algorithm”, The SIJ
Transactions on Computer Science Engineering And
Applications(CSEA), vol.1, March-April 2013.
[13] Bhise R.B, Thorat S.S and Supekar A.K, “Importance of data mining in
higher education system”, IOSR Journal of Humanities and Social
Science, ISSN: 2279-0837, vol.6, iss 6, January-February 2013.
[14] V.Ramesh, P.Parkavi and P.Yasodha, “Performance analysis of aata
mining techniques for placement chance prediction”, International
Journal of Scientific and Egineering Research , Vol.2, iss 8, August
2011.
[15] Mohammed M.Abu Tair and Alaa M.El-Halees, ‘Mining Educational
Data to Improve Students’ Performance: A case Study’, International
journal of information and Communication Technology Research,
ISSN: 2223-4985, vol.2 no.2, February 2012.