INTERNATIONAL JOURNAL OF PROGRESSIVE RESEARCH IN SCIENCE AND ENGINEERING, VOL.4, NO.05, MAY 2023.

College Recommendation System Using Machine Learning

Rajaram Ambole 1, Kartiki Hinge 1, Pallavi Jatiya 1, Renuka Bhosale 1, Shrddha Bolshette 1
1Student, Department of Computer Engineering, Vidya Pratishthan’s Kamalnayan Bajaj Institute of Engineering & Technology, Savitribai
Phule Pune University, Baramati, Pune, India.
Corresponding Author: [email protected]

Abstract: - Now-a-days selecting an appropriate college is a big deal for students. Many students face problem while selecting the
proper college which will fit their requirements. So, using machine learning college recommendation systems are made to assist
students, guide and recommend the perfectly suitable and best colleges to the students. This system is ‘College Recommendation
System’ which will help students completing their HSC and willing to take admission in best engineering college according to their
marks and requirements. Algorithms used in this system are machine learning algorithms as like semi supervised learning algorithms,
decision trees, etc. Building such a system for students will reduce the student’s time in searching the colleges. This application
requires the user to fill in some of the mandatory details and provides the student with the probability of securing admission in the
college.

Key Words: - College Recommendation, machine learning, semi supervised learning algorithms.

I. INTRODUCTION For using this system, the student, he/she has to do the
registration and create a profile. While creating the profile user
College Recommendation System is a web application for must enter the appropriate and valid information about himself
students to make the college selection procedure easier. This /herself. For creating the profile user has to enter the name,
system will help the students to choose the best among huge marks obtained in HSC and entrance exams for which he had
number of options. So, this system works well by considering appeared, his 12th std. college name, the preferred University
all the parameters like marks obtained in HSC, entrance exams, etc. This system considers all those parameters and recommend
university they want to prefer. Depending upon the users input a college for students, so that it becomes easier for them to fill
the system will work and recommend a college which fits the forms for college selection and get the seat in best college
his/her requirements. So, developing such a recommendation in accordance with the marks and requirements. In addition to
system is a big deal as it needs the bigger dataset but it’s quite the recommendation this system also provides the list of
interesting to work with this project. This system is most colleges and its information so that it will get easier for students
needed system because this system will reduce the student’s to compare and analyze the colleges and select among it. The
burden and work for searching and analysing the colleges. So, list will be displayed on the home page where the user will be
using this it will be much affordable and easier to get the able to see the abstract information about the college. So, such
relevant college list. This system will be free for students and type of system is very helpful and beneficial to the users which
hence it will be easy for them to use it and get help from us. will assist them to select the college within few clicks.

II. LITERATURE REVIEW

Manuscript revised May 28, 2023; accepted User friendly college recommendation system was specially
May 29, 2023. Date of publication May 31,
2023. made for the 12th completed students. This system uses user’s
This paper available online at www.ijprse.com
Facebook profile for taking the information. In their system
they have developed a recommender system that understands
ISSN (Online): 2582-7898; SJIF: 5.59
user’s need and accordingly generates recommendations for
him/her through simple inter- face, which provides information
to the students and advisors to improve the choice of courses.

RAJARAM AMBOLE., et.al.: COLLEGE RECOMMENDATION SYSTEM USING MACHINE LEARNING 313
 INTERNATIONAL JOURNAL OF PROGRESSIVE RESEARCH IN SCIENCE AND ENGINEERING, VOL.4, NO.05, MAY 2023.

They have used hybrid approach which combines advantages similarity to predict the best-matched subject and colleges for
of Collaborative filtering algorithm and Content based filtering the profile of the user. [2]
algorithm to improve our recommendations. Their system
consists of an interactive user profiling process in order to The system contains variety of range of filters through which
construct profile of the User. The system consists of no of student can get list of colleges based on his/her criteria. Other
attributes based on which questions are framed. This system system build on same objective are web page sites while our
also shows college rankings. Algorithms used in this system, system is desktop application that a user can use with ease. This
Collaborative Filtering, Content Based Recommendations, proposed system aims to develop the software for SSC passed
Demographic Recommendations, Knowledge Based students in order to help them in their admission process. This
Recommendations, Hybrid Recommendations. This system project helped the high school students to get the list of colleges
provides one stop portal as the source of information for all based on their scores. Algorithms used Naive Bayesian’s and
academic institutions. But extracting Knowledge from the Decision Trees which are useful to solve the given problem.
generated user profile was a big deal. [1] The algorithm which has higher accuracy in recommending the
best college is to be used. Therefore, this technique would be
helpful to students for minimizing their time in searching
colleges. Their system consists of total four modules which
describe the various aspects for recommending colleges, de-
tails of the all colleges, branches and comparison with other
colleges. [3].

The algorithm has been developed can be modified accordingly

so that it will function properly for the new pattern as well, for
obtaining admission for different colleges. The main focus here
has been given to the engineering field and the data has been
collected accordingly. So, the students opting for the
Designing a recommendation system that could understand the engineering fields may enter their marks in order to get an
skill set and interest of a student through the data from the appropriate for the colleges suitable for them. Similarly, this
user’s profile to suggest recommended options of colleges for system is used for several other fields such as medical,
the users to select. This is a college recommendation system pharmacy etc. The system is designed for two modules Admin
which takes into account all the major factors on which the and Student respectively, which are the basic working areas in
choice of the right college depends. The system also helps the the system and which have limited modules in it. In the College
students in selecting the right college for them as important Admission Predictor and Smart List Generator the developer
factor such as score, preferred study subjects / domain of study, used Ada boost algorithm. The Ada Boost algorithm is a binary
areas and countries and also other parameters are considered for classifier, which classifies the given input into two classes. It
presenting the best matches. Also, they have some additional classifies the input data based on the classification from
options for the students who want to pursue P.H.D as these multiple decision trees. The majority of the classification
students want to pursue their degree under some professor. obtained from the determined number of decision trees is the
They also provide the students with the most matched faculties classification of the algorithm.
with the student’s profile. This also helps students in
III. PROPOSED MODEL
constraining their search criteria. System’s react server calls the
LinkedIn API which redirects the user and asks permission
The system is designed so that it will help students to get the
from the user to allow our app do authentication through his
most appropriate college in accordance with their marks
LinkedIn account also to use his basic information to provide
obtained in 12th, CET, JEE, and other preferences given by
accurate results in the app functions. The access token is used
user. The system will also display college lists with its
to fetch user details and basic LinkedIn user profile from the
information. The algorithms used in performing the crucial task
LinkedIn API. It applies TF-IDF vectorization and cosine

RAJARAM AMBOLE., ET.AL.: COLLEGE RECOMMENDATION SYSTEM USING MACHINE LEARNING 314
 INTERNATIONAL JOURNAL OF PROGRESSIVE RESEARCH IN SCIENCE AND ENGINEERING, VOL.4, NO.05, MAY 2023.

of recommendation i.e prediction of results is done using • Logistic Regression- Logistic regression is an easy
machine learning algorithms like random forest and decision algorithm that estimates the association between one
tree. In this way the system is designed by considering all the dependent binary variable and independent variables,
demerits and limitation of previously designed systems. Further computing the probability of the occurence of an
we can work on all these things to make the system more event.
reliable and effective.
3.3 Architecture
3.1 Process System Architecture is designing the system with front-end,
In first stage of the project, we will start making frontend and back-end, module design, pre-processing of data, training and
back-end of the system. Then we started preparing the database testing the module, and develop the recommendation system,
which will contain the list of all Engineering colleges, user’s that would predict the college for given input data.
details, registration details, cut-off details, login details and
many more useful information. All information of colleges that
we have put in database is be based on student’s reviews. These
reviews are collected by us personally from some of the
students of various colleges and by the way of google sheets.
These reviews are be based on various parameters like teaching,
college environment, training, etc. Our system compares
students score with college’s cut-off and generate list. System
also generates ratings based on users provided reviews. Our
system, predict colleges using efficient prediction algorithm.
Datasets used in building this model are created and made by
us, we have gone through different college cutoffs, understood
their criteria and done the dataset entries. Fig.1. System Architecture

3.2 Algorithms 3.4 Features

• Decision Tree- Decision tree is used to find the class The features of the system maintain the students records in
of dataset. There is a method known as attribute limited time and less paper work, which helps the students to
selection measure that may be used to choose the predict more accurate college in limited efforts.
Table.1. The Features of the system.
optimal attribute for the root node which is based on
Information Gain and Gini Index.
• Random Forest- Random Forest is a classifier that
uses many decision trees on different subsets of the
provided dataset and averages the results to increase
the dataset’s predicted accuracy. Instead of depending
on a single decision tree, the random forest uses
forecasts from each tree and predicts the result based
on the votes of the majority of predictions. Using the
training set that best matches the trees for these
samples and the training data (p = p1, q = q1,.....pn)
with answers (Q = q1,.....pn) and bagging (X times),
select a random sample and replace it as follows: For
x = 1. . . , X:
• Support Vector Machine- The SVM algorithm’s
objective is to establish the best line or decision
boundary that can divide n-t

RAJARAM AMBOLE., ET.AL.: COLLEGE RECOMMENDATION SYSTEM USING MACHINE LEARNING 315
 INTERNATIONAL JOURNAL OF PROGRESSIVE RESEARCH IN SCIENCE AND ENGINEERING, VOL.4, NO.05, MAY 2023.

IV. RESULTS AND PERFORMANCE ANALYSIS Table.2. Parameters Accuracy and Time complexity

Our proposed system works as follows; Admin: firstly, admin

can add, update, delete add the dataset and view dataset and also
view complaint/feedback/suggestion.
User can register and login by using registered mail ID and get
the list of recommended colleges.

V. CONCLUSION

In the proposed system, we have implemented machine

learning algorithms to predict the college. At the end our model
can predict college to a student based on the criteria given. We
can use this model for various other fields like medical colleges,
PHD colleges, junior colleges, diploma colleges. This is very
useful for students to make the right choice while selecting their
college for further studies. This model gives advantages to
students so that the college selection procedure becomes easier
and within few clicks they can get their expected outcome. So
this system is better option for students completing their HSC
and looking for institutes to take admission in college for their
degree. A college recommendation system’s top aim is to
accurately predict the colleges for students so that they can
visualize their eligibility to take admission at their dream
university or college. The efficiency of the system for
predicting the results depends on the parameters and the
features, volume of dataset given for training the module. The
machine learning algorithms used in prediction model are
decision tree and random forest, these algorithms help to give
simpler and appropriate predictions using the larger dataset
considering quality features. So, these machine learning models
worked efficiently to give accurate predictions based on given
data.

REFERENCES
[1]. Aryan Saxena; Prabhangad Singh; Shailendra Narayan
Singh, Gender and Age detection using Deep Learning,
IEEE, 2021.
[2]. A User-Friendly College Recommending System Using
User-Profiling and Matrix Factorization Technique. (May
2017).
[3]. Vidish Sharma, Tarun Trehan, Rahul Chanana, Suma Dawn,
“StudieMe: COLLEGE RECOMMENDATION SYSTEM”
(2019 3rd International Conference on Recent
Developments in Control, Automation Power Engineering
(RDCAPE)).

RAJARAM AMBOLE., ET.AL.: COLLEGE RECOMMENDATION SYSTEM USING MACHINE LEARNING 316
 INTERNATIONAL JOURNAL OF PROGRESSIVE RESEARCH IN SCIENCE AND ENGINEERING, VOL.4, NO.05, MAY 2023.

[4]. Vinit Jain, Mohak Gupta, Jenish Kevadia, Prof. Krishnanjali

Shinde, “College Recommendation System”. (International
Journal of Engineering Research Technology (IJERT) ISSN:
2278-0181 Published by, www.ijert.org ICIATE - 2017
Conference Proceedings, Special Issue – 2017).
[5]. Jadhav Sayali Pramod, Patil Durga Ananda, Kamble Sonali
Arun, Shinde Sumit Sahebrao, “College Recommendation
System using Machine Learning” (International Research
Journal of Engineering and Technology (IRJET) e-ISSN:
2395-0056 Volume: 07 Issue: 03 — Mar 2020 www.irjet.net
p-ISSN: 2395-0072).
[6]. Kiran Kumari (Assistant Professor), Meet Kataria, Viral
Limbani and Rahul Soni [Students], Department of
Information Technology, “College Admission Predictor and
Smart List Generator”, K. J. Somaiy.
[7]. Annam Mallikharjuna Roa, Nagineni Dharani, A. Satya
Raghava, J. Buvanambigai [Students], K. Sathish (Assistant
Professor), Computer Science and Engineering,” College
Admission Predictor (Volume No. 4, April 2018)”, SRM
Institute of Science and Technology, Chennai, Tamil Nadu,
India, (JNCET) www.jncet.org Volume 8, Issue 4, April
(2018).

RAJARAM AMBOLE., ET.AL.: COLLEGE RECOMMENDATION SYSTEM USING MACHINE LEARNING 317