See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/258652670

Prediction of Final Result and Placement of Students using Classification

Algorithm

Article in International Journal of Computer Applications · October 2012

DOI: 10.5120/8945-3111

CITATIONS READS
18 1,620

2 authors:

Neelam Naik Seema Purohit

Usha Pravin Gandhi College of Arts, Science and Commerce, VileParle, Mumbai Brihan Maharashtra College of Commerce-BMCC
5 PUBLICATIONS 28 CITATIONS 30 PUBLICATIONS 99 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Project on credit evaluation for Banks View project

All content following this page was uploaded by Seema Purohit on 06 May 2016.

The user has requested enhancement of the downloaded file.

 International Journal of Computer Applications (0975 – 8887)
Volume 56– No.12, October 2012

Prediction of Final Result and Placement of Students

using Classification Algorithm
Neelam Naik Seema Purohit
Assistant Professor Director, NMITD
TIMSCDR, Kandivli Dadar, Mumbai
Mumbai, Maharashtra, India Maharashtra, India

ABSTRACT improving managerial skills of candidate. Masters in Computer

The quality higher education is required for growth and Applications (MCA) course provides professional computer
development of country. Professional education is one of the technological education to students. This course provides state
pillars of higher education. Data mining techniques aim to of the art theoretical as well as practical knowledge related to
discover hidden knowledge in existing educational data, predict information technology and make students eligible to stand in
future trends and use it for betterment of higher educational progressing information industry.
institutes as well as students. The objective of this study is to use No system has been developed to predict result of MCA students
prediction technique using data mining for producing knowledge or to predict placement of student in any company, before
about students of Masters of Computer Application course admitting him or her in an institute. The prediction of MCA
before admitting them to the course. result or the prediction of category of company where student
can be placed after the completion of MCA course will help to
General Terms channelize efforts of students for proper progress. It will also
Data Mining help teachers to take corrective measures towards the progress
of the student during the course. It will also help institute
management to admit those candidates who will be placed in
Keywords large scale organizations after completion of the course work. It
Data Mining, Higher education, Classification, Prediction will help to build reputation of institute in existing similar
category institutes in terms of IT education provider and creator
1. INTRODUCTION of well versatile and skilled employees for IT industry.
Education system is backbone of progressing Indian society. The present study concentrates on the prediction of results and
India’s superiority in information technology is due to evolution placements of MCA students when the student is admitted to the
in higher education system in last few years. One has to course.
concentrate on challenges faced by higher educational institutes
to improve quality of higher education. Pillania explains that 2. DATA MINING AS A POWERFUL
there is lack of research culture in higher education institutes for TOOL
generating knowledge and lack of interaction between industry Data mining is applicable to all sectors of business, like retail,
and higher education institutes [4]. The industry interaction banking, telecommunications, marketing, web mining,
would help higher education institutes for development of medicines and many more. Data mining is widely used in
curriculum according to the industry requirements, providing business context but its applications in academics are very much
hands on experience for students for applying theoretical limited. Amar Sahay has developed a software system to assist
knowledge. World Bank report (2010) concentrates on rapidly higher education in assessing and predicting key issues related
changing technology needs retraining of workforce [8]. to student success such as enrollment management, dropout rate,
Bhattacharya emphasizes on requirement of repository of latest time to degree, and suggests ways to improve courses and
research in the topics related to India specific needs [2]. The programs [1]. Singleton Alex utilized national database,
repository should be accessible to students, teachers, examined systematic grouping of interest within course choice
researchers, curriculum developers in higher education sets and assess whether certain types of student are more likely
institutes. According to Rashan, there is a need of improvement to make concentrated or diffuse subject selections [7]. This
in education system by promoting understanding of individual creates an information base of generalized course choice
learner capabilities by encouraging problem solving skills, behaviors that higher education institutions could utilize for
critical thinking and personalized learning [6]. He emphasizes targeted marketing, recruitment and selection activities, and
on data mining techniques to gather extra information about additionally forms the basis of decision support framework that
academic progress of each student in advance and to help could be implemented in a variety of online tools to help
teachers to manage extra coaching and to provide proactive students. Ranjan Jayanthi et al. propose a framework for an
feedback to the student. effective educational process using data-mining techniques to
Higher Education is categorized into professional and non- uncover the hidden trends and patterns and making accuracy
professional education. Professional education provides based predictions through a higher level of analytical
professional knowledge to students so that they can make their sophistication in the process of counseling students [5]. Jing
stand in corporate sector. Professional education may be Luan addresses the capabilities of data mining and its
technology oriented or it may be totally concentrating on applications in higher education [3]. Three case studies of J.

35
 International Journal of Computer Applications (0975 – 8887)
Volume 56– No.12, October 2012

Luan demonstrate how data mining saves resources while examination, if his trade of graduation is mathematics, physics,
maximizing efficiency, and increases productivity without information technology or computer science.
increasing cost.
5. DATA COLLECTION
3. OBJECTIVE OF THE STDUY This study is based on the documents showing previous
As per the literature review in section 2 earlier works done on academic record of students of MCA institutes in Mumbai
applicability of data mining to handle different issues in higher (India). The sample comprises around 325 students from various
education, but do not focus on the predictability of the MCA institutes in Mumbai. The each record of student contains
placement of the students corresponding to the different Gender, SSCE (Secondary School Certificate Examination)
parameters that are responsible for making predictions. The percentage, HSCE (Higher Secondary Certificate Examination)
enrollment and successful running of professional courses percentage, Graduation percentage, Graduation trade, MCA
depends upon the placements and the institutes running these result and details of organization where student has been placed.
courses would like to analyze the students’ data to find a Around 195 records are used to create model for prediction of
concrete association between the parameters affecting the result and placement. Remaining 130 records are used to
placement activity. validate developed model. Table 1 shows sample data. XLminer
Hence the research objective of the study is to use data tool is used to apply data mining algorithms and to develop
mining algorithms to predict final result and placements of classification tree. Decision rules are developed according to the
graduate student based on previous academic record. classification tree generated. These decision rules are
implemented in a software using dot net software platform.
4. HYPOTHESIS OF STUDY
Following hypotheses are made and verified through study. 6. DATA PROCESSING
H1. The final result of MCA student is always related to the Data from different sources must be converted into a common
previous academic record of the student such as SSC (Secondary format for processing. Some data may be encoded or
School Certification), HSC (Higher Secondary Certification), transformed into most suitable formats. Data reduction may be
Graduation percentage, and Graduation trade. used to reduce the number of possible data values being
H2. The final placement of MCA student either in large, considered. Various fields of records are converted into
medium or small scale organization after the completion of numerical form. Each of the trade of graduation from BCA
MCA course is predictable on the basis of the previous academic (Bachelor of Computer Application) to B.Sc.(Bachelor of
record of the student such as SSC, HSC, Graduation percentage, Science) is numbered from 1 to 7 on the basis of inclusion of IT
and Graduation trade. or computer science related syllabus in graduation curriculum.
H3. MCA students getting marks around 60% at undergraduate, The gender field values, 0 for male and 1 for female are also
graduate level, receives better employment if his graduation included. The output field is percentage of MCA result of every
trade is mathematics, physics, information technology or student, which is converted into categorical form and numbered
computer science. from 1 to 8. The students are placed in various companies like
H4. MCA students from commerce, zoology, chemistry small scale, medium scale or large scale. The categorization of
background scores more in MCA examination even if his around 30 companies is done on the basis of revenue of
undergraduate as well as graduate marks are less than or equal to company and number of employees working in the company.
60%.
H5. MCA Students getting marks around 60% at undergraduate
as well as graduate level receives better score in MCA

Table 1: Sample data showing MCA result of students

no NAME SSC HSC Gender GTRADE GPER TOTAL MCA% MCA%R
1 AWASTHI VARUN C. 73.60 65.33 0.00 4.00 60.57 2360.00 67.43 4.00
2 BAPAT OMKAR S. 80.00 67.83 0.00 9.00 65.43 2666.00 76.17 6.00
3 BATTE DEVANSHREE V. 70.26 64.17 1.00 1.00 64.04 2373.00 67.80 4.00
4 CHOUDHARY PANKAJ S. 66.00 61.00 0.00 3.00 76.45 2394.00 68.40 4.00
5 CHOURASIA RITASHA 61.60 63.17 1.00 3.00 63.68 2120.00 60.57 3.00
6 CHOURASIYA VISHAL B. 69.73 61.83 0.00 5.00 52.13 2220.00 63.43 3.00
7 DESAI MEGHA A. 84.26 80.67 1.00 1.00 74.33 2699.00 77.11 6.00
8 DESHPANDE NEHA S. 72.20 66.80 1.00 2.00 72.25 2361.00 67.46 4.00
9 DIVITY SHRAVAN V. 72.40 62.67 0.00 6.00 64.50 2411.00 68.89 4.00
10 D'SOUZA VERON 61.86 56.33 0.00 3.00 73.80 2221.00 63.46 3.00
11 DUBEY CHANDRASHEKHAR L.
73.60 56.33 0.00 4.00 51.00 2234.00 63.83 3.00
12 GADDAM SRIMANTHKUMAR 65.06
D. 44.00 0.00 5.00 53.13 2049.00 58.54 2.00
13 GORE RAHUL R. 82.66 69.17 0.00 3.00 60.06 2110.00 60.29 3.00
14 GUDKA KALPESH R. 51.86 56.83 0.00 3.00 77.23 2617.00 74.77 5.00
15 JAIN BHAVIK P. 73.33 53.67 0.00 4.00 54.37 2169.00 61.97 3.00

36
 International Journal of Computer Applications (0975 – 8887)
Volume 56– No.12, October 2012

7. DATA ANALYSIS USING DATA the prediction of final result and placement of students are
shown in figure 1 and figure 2 respectively.
MINING
XLMiner 3.2.8 version is used for applying classification
algorithm on given data. The generated classification trees for

57.08
40 5
59

59.62 HSC
67.5
15 5 25
38 21

65.53 GRAPER 5.5 60.66 HSC 72.79

5 5 5
20 12
9 6 7 14
SSC GTRADE HSC 26 SSC
53.65 3 58.75
3 2 5 3
81.99
5
7 5
13 HSC 7 HSC 6
58.98 8
3 2 SSC
2 19
65.20 7 4 5
5 GRAPER
5
8
3 63.83
GRAPER 5
11
2 2 8
HSC

64.52
3
6
5
GRAPER

4 2

Figure 1: Classification Tree for MCA result- Full Tree (Using Training Data)

37
 International Journal of Computer Applications (0975 – 8887)
Volume 56– No.12, October 2012

66.96
41 5
58

SSC

62.34 45

20 21
GRAPER GTRADE 16
42

56.58
52.96 63.83 62.94
5 5
8
12 HSC 11 23
HSC 7 9
2 10
19 GRAPER

56.62 HSC
5 61.99 2 3
5 60.66
7 5
5
6
GRAPER SSC 5
HSC 60.31
68.95 5
0 1 5
3 2 18
5
6 GRAPER
5
5 GRAPER
13 80.79
0
5
1 2 71.8 1
13 5
SSC
6 7
SSC 1 1
1 2

Figure 2: Classification Tree for Placement Prediction - Full Tree (Using Training Data)

Decision rules are developed based on these trees. ASP.net commerce ) will secure first class even if during previous
software is used to develop web forms to implement these rules. academic career student have not secured first class. This
Table 2 shows results after testing records on developed tool. In supports hypothesis 1 that is the final result of MCA student is
76.19% cases students (graduation trade: Physics, Mathematics, always related to the previous academic record of the student
computer science and information technology) and in 65.38 such as SSC, HSC, Graduation percentage, and Graduation
cases students (graduation trade: Zoology, Chemistry and trade.

38
 International Journal of Computer Applications (0975 – 8887)
Volume 56– No.12, October 2012

Table 2: Effect of academic record on MCA result and placement of student

MCA Result Placement

Graduation SSC HSC GRA
Trade Result Result Result
< 60% >= 60% No Yes

Physics, Maths,
<= 60% <= 60% <= 60% 23.80% 76.19% 50.00% 50.00%
CS, IT

Zoology,
Chemistry, <= 60% <= 60% <= 60% 34.61% 65.38% 42.31% 57.69%
Commerce

Table 2 shows that in 50% cases students (graduation examination even if his undergraduate as well as graduate marks
trade: Physics, Mathematics, computer science and information are less than or equal to 60%.
technology) and in 57.69 cases students (graduation trade: Table 2 shows that if the student is from physics,
Zoology, Chemistry and commerce) will be placed in large scale mathematics, computer science or information technology
organization even if during previous academic career student background and he has marks around 60% at undergraduate
does not have secured first class. This supports hypothesis 2 that level as well as graduate level then chances of getting first class
is final placement of MCA student either in large, medium or in MCA examination is 76%. As the graduation trade is
small scale organization after the completion of MCA course is computer based or mathematical based, they can score most at
predictable on the basis of the previous academic record of the MCA level even though their SSC, HSC and Graduation
student such as SSC, HSC, Graduation percentage, and percentage is less than 60. This supports hypothesis 5 that is
Graduation trade. MCA Students getting marks around 60% at undergraduate as
well as graduate level, receives better score in MCA
If the student is from physics, mathematics, computer examination, if his trade of graduation is mathematics, physics,
science or information technology background and he has information technology or computer science.
secured marks around 60% at undergraduate level as well as
graduate level then chances of getting placed in large 8. CONCLUSION
organization is 50%. This is because as the graduation trade is The created classification trees for MCA result prediction
computer based or mathematical based, they can be technically and placement prediction of students are tested for validation
sound, thorough with their basic concepts and can face data. The overall error occurred to classify validation data using
placement interview successfully. This supports hypothesis 3 MCA result prediction classification tree is 38.46% while for
that is MCA students getting marks around 60% at validating placement prediction classification tree it is 45.38%.
undergraduate, graduate as well as post graduate level, receives It indicates that final result and placement of students can be
better employment if his graduation trade is mathematics, predicted before hand admitting them to the course.
physics, information technology or computer science.
The distinctive observation made through present study is
Distinctive features of the research work done as shown in that, both the classification trees generated do not take gender as
Table 2 indicates that if the student is from commerce, zoology an attribute for taking decision at any level. It indicates that final
or chemistry background and his under-graduation and result and placement prediction of MCA students is not gender
graduation percentage is around 60 then the chances of scoring specific. For few records it is found that even though academic
first class at MCA exam is 65%. The reasons for this in the record of student is not satisfactory and he/she is not from
opinion of the researcher are two; one is that as MCA course information technology background, placement of such student
contains approximately 50% theoretical subjects, students from has been done in large scale organization. The reason behind
commerce and non-mathematical science background can score existence of such outliers is the student’s skill to face interview
more marks in MCA examination because in theory subjects smartly, his/her confidence level and proper behaviour during
they can score more than the practical subjects even if they have interview. The knowledge gained through above analysis can be
less marks in their graduation courses and the other is that the utilized for further enhancement of students; institutes and it can
mindset of such students is to study hard from the starting of the provide feedback for improvement of teaching by faculty
course as they know that they are from not-IT background. members of the higher education institutes.
This supports hypothesis 4 that is MCA students from
commerce, zoology, chemistry background scores more in MCA

39
 International Journal of Computer Applications (0975 – 8887)
Volume 56– No.12, October 2012

9. REFERENCES [5] Ranjan Jayanthi and Malik Kamna 2007, “Effective

educational process: a data-mining approach”, The journal
of knowledge management systems, vol. 37, No. 4
[1] Amar Sahay 2010, “Assisting Higher Education in
Assessing, Predicting, and Managing Issues Related to [6] Rashan K.H. 2010, “Data Mining Applications in the
Student Success: A Web-based Software using Data education sector”, MSIT, Carnegie Mellon University
Mining and Quality Function Deployment”, Academic and
[7] Singleton Alex 2009, “Data Mining course choice sets and
Business Research Institute Conference, Las Vegas
behaviours for target marketing of higher education”,
[2] Bhattacharya 2005. “Changing Strategies in Professional Journal of Targeting, Measurement and Analysis for
Education” By Prof. S.K. Bhattacharya in IETE technical Marketing Vol. 17, 3, 157–170
review, vol. 22, no. 5, September
[8] World Bank Report 2010,
[3] Jing Luan 2004, “Data Mining Applications in Higher “http://www.worldbank.org.in/WBSITE/EXTERNAL/CO
Education”, SPSS Executive Report UNTRIES/SOUTHASIAEXT/INDIAEXTN/0,,contentMD
K:21493265~pagePK:141137~piPK:141127~theSitePK:29
[4] Pillania Rajesh K. 2007, “Knowledge Management for 5584,00.html”, accessed on June 15, 2010
Indian Business Schools”, Journal of Services Research,
Vol. 7, No. 2

40

View publication stats