A WEB BASED APPLICATION FOR TUTORING SUPPORT IN HIGHER

EDUCATION USING EDUCATIONAL DATA MINING

Suvarnalata Hadge1,Bhavika Patil 2, Swapnali Ahire 3,Pushkar Jadhav 4, S.G.Bodake5,
G1234PG Student, Dept. of Information Engineering, Sandip Polytechnique, Maharashtra, India
5Assistant Professor, Dept. of Information Engineering, Sandip polytechnique, Maharashtra, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract – This Paper aimed at developing on Online admission due to they did not able to reserve the seat
College Prediction System that is of importance to the because of make mistakes in college selection. This type of
college as well as students. In recent time for student to condition makes additional threats to the student career.
select college for higher education is very difficult as per
their aggregate percentage. After the complete of diploma To avoid this type of condition, student required some
for higher education admission limited seats are available. helping platform which gives them correct knowledge or
Student don’t know for his/her which college is applicable. information about college. In this context, this study
This study focuses on ways to support student in focuses on supporting Students in making admissions
admissions decision making using data mining techniques decisions through the application of data mining
to predict colleges based on previous cut-off performance techniques to better predict College before designation.
at institute. In college prediction system based on valid
This would support institute decision makers as they set
and reliable cut- off criteria is very important to select
efficient admissions principles. This is because they use
colleges for higher education. Every college enroll their
only formal statistical methods rather than new and
college in college module then admin have a right to
efficient predictive techniques such as Educational Data
approve or disapprove if the college is centralized then
Mining (EDM), which is the most popular technique to
admin approved this college if college decentralized then
evaluate and predict college cut-off -performance. EDM is
admin disapprove this college. centralized college is show
the process of extracting useful details and models from a
to student as per their percentage
huge educational database, which can then be used to
Key Words: Data mining techniques1, Educational data predict students’ performance.
mining2, College Cut-off Prediction4, Online System5,
Database6, Student Aggregate7. 1.2 System Design

We are having a three module which is Admin Module,

1. INTRODUCTION College Module and Student Module. Number of institutes
register on college module and student also register on
In Recent Time, all higher education institutions, module in this database huge number of records are
especially diploma pass out student’s, face challenges in stored. Admin manage the institute but the user is self-
the admission process. Each university should strive for an approved. System is process on this institute and student
admissions system based on valid and reliable admissions data and give the perfect prediction from the data.in
criteria that select student likely to succeed in its cut-off institute data includes the institute Name, Code,
criteria. In addition, each student should use the best Designation, Institute Mobile no, Institute Phone No,
possible techniques for predicting college previous cut-off Email, District, Taluka, Address of Institute and Password.
criteria for admitting them. In student data contains Gender, Cate, Name, Enrolment
No, Percentage, Passing Year, Institute name, University
Each student faces some issues because in admission system,
Name, Student Mobile No, Email, Department, State, Taluka
limited seats are available that’s why student's get confuse
and Password. On the basis of this data, system is done the
because in India each college have a different category
prediction using data mining Technique. Data mining is
criteria. In category wise limited seats allocated for each
one of the most useful procedures that help entrepreneurs,
college, then student gets confused and put wrong option
researchers, and individuals to extract valuable
according their aggregate without seeing criteria and then
information from huge sets of data. Data mining is also
unfortunately lost all this college options. Second way of
called Understanding Discovery in Database (KDD). Using
admission is management through but in that also college
the data mining technique institute get shortlisted
requirement is very high for fees, every student is not
candidate for admission process.
affording these fees and student take one year gap and
student loss their one year. On a survey basis there are
80% to 90% student did not find perfect college and lost
their

| | |
 In this study, different admissions criteria were used as
input attributes to predict student’s second-year aggregate
in the computer sciences colleges of engineering.

Most of the few published related studies performed in

India have been confined to engineering colleges. These
studies have not used EDM, which can discover hidden
patterns in institution’s large datasets, and therefore
enhance their decision making. One of the very few studies
related to this topic that has been confined to a computer
science college is study, which applied one of the EDM
techniques to predict student’s final aggregate based on
their grades in all semester. The authors of collected data
Fig-1: System Architecture from the institute of the Computer Science year wise cut-
off. They identified which courses in the study plan most
2. Literature Survey strongly affect for direct-second year student. However,
they applied only one EDM technique on a very small
Several data higher learning institutions, student database to predict college above their final aggregate
percentage is the factor most important to a university’s percentage in diploma and did not ensure the accuracy of
quality. EDM is currently the technique most commonly used their results by using additional EDM techniques.
by researchers to evaluate and predict student
Another study that related to this topic is study of each
performance due to its significance in decision making.
institute because each institute has different cut-off of
Predicting college cut-off to student aggregate to has two each year as branch wise and category wise, which applied
main factors: attributes and prediction methods. It has key based algorithm to search student aggregate based on
been shown in that student aggregate is the most frequently their college criteria above their last year cut-off of college.
used attribute in predicting student performance at By comparing results of the college show above their final
university. It has been used in many examinations. Other year aggregate using prediction techniques, they found
attributes commonly used by exploration to predict that key based search outperformed others by achieving a
college cut-off by student percentage at university are: prediction accuracy rate of 99%.
category wise and branch wise. Few researches have used
However, the studies have not assessed the relationship
other attributes like extra student demography, and college
between admission criteria and computer science student
information.
performance in universities, which is the focus of this
However, input variables such as college information are study.
used to predict colleges cut-off to student aggregate of
To the best of our knowledge, the only related published
diploma for university in the college prediction process for
study that addressed this topic is study. In this study, the
student higher education admission. This is the focus of this
authors applied classification techniques on a database of
study. Several data mining classification techniques have
college to determine whether pre-result of cap round
been applied for college prediction for student admission.
process have a significant effect on student aggregate.
For example, The institute allocated seat for minimum 20
They found that institute final grade affects college more
and student apply their seat is 500 extra, so student don’t
than student. These findings differ from the findings of this
know which college applicable above their marks and
study.
student have only three round for college if student select
college, but their mark student not any seat allot then However, their data was collected from all colleges, all
student applicable for management quota if student department (Computer, Civil, Electrical, E&TC,
financial is poor then student not admit for college so it Mechanical) and all category (Open, OBC, SC, ST, NT) of the
gates year drop means student loss their one year because College of higher education at India, which has five
of their mistake to fill form wrong, if student fill right departments. In addition, they only used one key based
college above their category wise criteria follow and search algorithm to search the key as department wise,
previous cut-off then student don’t have any loss so we category wise, and all other fields match on their data
study this process. rather than applying different EDM models to predict
colleges above colleges, which could have enhanced their
The analysis of the literature reveals that most available
findings.
studies do not explore the relationship between college
cut- off scores and their expected college in the admissions
3.1 Algorithm
process. Thus, this study fills a research gap.

| | |
Input:
root, trapdoor containing keyword to be searched.

| | |
Output:
3.2 Flowchart
pointer to the documents containing the keywords; NULL
if non-exist.

1. Start

2. NODE_x = Disk Read (root).

3. if NODE_x is an index node

(a) If there is an object o in NODE_x such that o: key

= keyword, return o: value.

(b) Find the child pointer x: child [I] whose key

range contains key.

(c) Return Search Query (NODE_x: child [I], key).

4. else If there is an object o in NODE_x such that o:key =

keyword, return o:value. Otherwise, return NULL.

5.end if.

The takes trapdoor and root as input and searches for the
keywords match in database.

The disk Read reads the corresponding root page from

disk to memory and returns the location in memory that
gets stored in node NODE_x. Fig 2-: Flowchart of College Prediction System
If NODE_x is index node then trapdoor is checked to see 4. Result
for keyword match. If found returns the corresponding
document pointed by the node. Otherwise based on
keyword, search will move to the child of NODE_x using Step1-: First go on browser there are three module
pointers. Admin, College, User each one has unique password and
unique username. Ech user has firstly Register herself then
The search continues recursively. Otherwise, if NODE_x login by username and password and Each college has
represents leaf then return the pointer to document if register herself then admin have authority to decide
search succeeds otherwise NULL. college has centralized or decentralized, the college is
centralized then admin approved the college otherwise
disapprove college. College is approved then college login
by username and password.

Fig3-: Home Page

| | |
Step2-: Go on college module and register the college. In Step5-: Check the college for admission which is perfect
that form all field are required. for you and you can take the admission in this colleges.

Fig 4-: College Registration form

Step3-: After register the college, if the college is Fig5-: Exact College match to Student Profile
centralized this is approved by admin and then we are able
to login otherwise it will give you error. After login college 5. CONCLUSIONS
can add the cut-off of the year and view the student which
applicable for the college. The aim of this study is to support Students in making
good decisions in its admissions process by predicting
College before filling the admission form. The study was
conducted with a largest dataset of college. The technique
used, however, are generally and can be used in any higher
education institution. The study confirms the effectiveness
of prediction College for higher education where decision
makers can use these models in planning and optimizing
colleges. In addition, the results show that a high-
performance model to predict colleges early performance
could be developed based on pre-admission information.
Also, college and student search using deep searching
means search with one or more keywords.

6. ACKNOWLEDGEMENT
Fig-5: Centralized College list Admin panel
With honor sense of gratefulness, we'd like to thanks all
Step4-: As a student go to the student module and the people who have lit our path with their kind guidance.
register, student is self-approved. Register using user We're thankful to these studious who did their best to help
name and password. during our project work. It's our proud honor to express
deep sense of gratefulness to P.M.dharmadhikari, Principal
of Sandip polytechnique Nashik, for his remarkable and
kind cooperative guidance to complete this project. We
remain grateful to Prof.V.B.Ohol, H.O.D. of Information
Technology Department for his timely suggestion and
precious guidance. The special appreciation goes to Prof.
S.G. Bodake sir excellent and precious guidance in
completion of this work. We thanks to all the associates for
their observable help for our working project. With
industry persons or lab technicians to help, it has been our
bid to throughout our work to cover the entire project
work. We also thankful to our parents who enhancing
their wishful support for our project completion
successfully. And we thanks to our all friends and the
Fig6-: Student Registration Form
people who are directly or indirectly associated to our

| | |
project work.

| | |
REFERENCES

[1] A. Dashputre, D. Shetye, A. Sonawane, P Mangalkar “A

web-based application for tutoring support in higher
education using educational data mining” Vol-8 Issue-2
2022.

[2] Hanan Abdullah Mengash, “Using Data Mining

Techniques to Predict Student Performance to Support
Decision Making in University Admission Systems” 30
March 2020, vol8 2020.

[3] H. Aldowah, H. Al-Samarraie, and W. M. Fauzy,

‘‘Educational data mining and learning analytics for 21st
century higher education: A review and synthesis,’’
Telematics Informat., vol. 37, pp. 13–49, Apr. 2019.

[4] A. M. Shahiri, W. Husain, and N. A. Rashid, ‘‘A review

on predicting Student’s performance using data mining
techniques,’’ Procedia Computer. Sci., vol. 72, pp. 414–422,
Jan. 2015.

[5] Prasanna B, C B Akki, “Dynamic Multi-Keyword

Ranked Searchable Security Algorithm Using CRSA and B-
Tree”, Vol. 6 (1), 2015.

[6] Dharni Negandhi, Aastha Mishra, Akshay Doshi,

“Online College Admission System”, Volume 5, Issue 2017.

[7] Jadhav Snehal Balasaheb, Supekar Bhagyashri Sitar

am, Wakode Vrushali Khushalrao, Vasaikar Nikita Ashok,
Mandlik Priyanka Bhausaheb “Web Based College
Admission System”, Vol.2, No.4, October

[8] 2011Surender Baswana, Partha Pratim Chakrabarti,

Sharat Chandran, Yashodhan Kanoria, Utkarsh Patange,
“Centralized Admissions for Engineering Colleges in India”
95, 2 (2005), 364–367.

| | |