COLLEGE RECOMMENDATION SYSTEM

Data preprocessing report submitted

in
partial fulfillment of requirement for the award of degree of

Bachelor of Technology
in
Information Technology

by

Mr. Parth Mandhare Mr. Prajesh Dhande

Mr. Prajwal Rathod

Institute Guide
Prof. Sonali Guhe

Department of Information Technology

G H Raisoni College of Engineering, Nagpur
(An Autonomous Institute affiliated to RashtrasantTukadojiMaharaj Nagpur University, Nagpur)
Accredited by NAAC with “A+” Grade
Ranked 130 by NIRF, MHRD in the Engineering Category for India Ranking 2021,
th

Ranked 2 by ARIIA 2020, MHRD in Private or Self Finance Institutions, 5 Star Rating by
nd

MIC, MHRD 2021

April 2023

1
 Declaration

We, hereby declare that the data preprocessing report titled “College Recommendation
System” submitted herein has been carried out by us towards partial fulfillment of requirement
for the award of Degree of Bachelor of Technology in Information Technology
Engineering. The work is original and has not been submitted earlier as a whole or in part for
the award of any degree / diploma at this or any other Institution / University.

We also hereby assign to G H Raisoni College of Engineering, Nagpur all rights under
copyright that may exist in and to the above work and any revised or expanded derivatives
works based on the work as mentioned. Other work copied from references, manuals etc. are
disclaimed.

Place: Nagpur Mr. Parth Mandhare

Date: 14-04-2023 Mr. Prajesh Dhande
Mr. Prajwal Rathod
(Name and Signature)

2
 Certificate

The Data Preprocessing report entitled as “COLLEGE RECOMMENDATION

SYSTEM” submitted by Parth Mandhare, Prajesh Dhande, Prajwal Rathod for the award
of Degree of Bachelor of Technology in Information Technology Engineering has been
carried out under our supervision. The work is comprehensive, complete and fit for
evaluation.

Institute Guide
Professor Sonali Guhe
Department of Information Technology Engineering
G H R C E, Nagpur

Head Director
Department of Information Technology Engineering
G H R C E, Nagpur GHRCE, Nagpur

3
 ACKNOWLEDGEMENT

We are personally indebted to a number of people who gave us their useful

insights to aid in our overall progress for this project. First of all, we would like to give
our deepest gratitude to our parents for permitting us to take up this course.
Our sincere thanks and heartfelt sense of gratitude goes to our respected
Director, Dr. Sachin Untawale our Head of Department, Dr. Mahendra Gaikwad for all
his efforts and administration in educating us in this premiere institution. We take this
opportunity to also thank our Project guide Prof. Sonali Guhe for her encouragement
throughout the project.
We also convey our sincere thanks to our Project Team members for their
efforts and hard work.

Thank You.

4
 LIST OF FIGURES

Sr. No. Figure No. Name of Figures Page No.

1 3.1 Inception modules where two 3 × 3

convolutions replace each 5 × 5 convolution.

2 3.2 Inception module that expands the filter

blanks while reducing the grid-size.

3 3.3 A 5-layer dense block with a growth rate of

k = 4.

4 3.4 Predictions made by our model

6 5.1 Importing the Required Python Libraries
7 5.2 Loading the Dataset into the System
Memory for Processing
8 5.3 Processing the Dataset

9 5.4(A,B) Analysing the Data

10 5.5 Split the Data into Training and Test Sets.

11 5.6(A,B) Preparing Image Data Generator Instance

From Keras For Image Augmentations
12 5.7(A,B) Preparing the DenseNet Model, Loading the
Weights and Compiling It
13 5.8(A,B) Implementing Checkpoint Settings and
Starting the Model Training
14 5.9 Saving The Trained Model as A HDF5 File

15 5.10(A,B) Testing the Model and Generating

Predictions.

16 5.11(A,B) App backend

17 5.11(C) Web Application Interface

18 6.1 Sample Dataset Images Interface

5
19 6.2 Training procedure of all the models

20 6.3 The Dense Block's Structure (k0 feature

maps on the first layer, k0+k feature maps

on the second layer, and k0+4k feature maps

on the third layer).

21 6.4 Model Comparison – Validation Accuracy &

Validation Loss
22 6.5 Results of the comparative study

23 6.6 Learning rate decay used in our study

24 6.7 Improved performance of the DenseNet

model after fine-tuning.

6
 LIST OF TABLES
Sr. No. Table No. Name of Tables Page No.
1 Dataset of Training the Model
2 6.1 Splitting dataset as peer the 60-20-20 rule
which means 60% of the images are in the
training set, 20% in validation set and 20 %
in the test.

3 6.2 Confusion Matrices (CMs) and performance

measures of different models

4 6.3 Model Accuracy and Loss of all models.

5 6.4 Total model parameter and trainable
parameter

6 6.5 Best model Verification by Time

7
 INDEX
Sr. No. Content Page No.
1. Introduction
1.1 Preface
1.2 Brief Overview
2. Literature Review
3. Methodology
3.1 Comparative Study
3.1.1 VGG 19
3.1.2 ResNet 50v2
3.1.3 Inception v3
3.1.4 Densenet 121
3.2 Evaluation
3.3 Fine-Tuning The Best Performing Model
3.4 Deployment
4. Data Collection
4.1 NIH Chest X-ray Dataset
4.2 Overview of Dataset
4.3 NIH Chest X-ray Dataset consist of 13 Common
Disease Categorie
4.4 Dataset consists of the following data
5. Modelling and Implementation
5.1 Importing the required python libraries
5.2 Loading the dataset into the system memory for
processing
5.3 Processing the dataset
5.4 Analysing the data
5.5 Split the data into training and test sets.

5.6 Preparing Image Data Generator Instance From

Keras For Image Augmentations

8
 5.7 Preparing the DenseNet Model, Loading the Weights
and Compiling It.
5.8 Implementing Checkpoint Settings and Starting the
Model Training.

5.9 Saving the trained model as a HDF5 File.

5.10 Testing the Model and generating predictions.
5.11 Website Interface SCANET
6. Testing & Summary of Results
6.1 Data Pre-Processing
6.2 Training Model Customization
6.2.1 VGGNet-19
6.2.2 ResNet 50 v2
6.2.3 Inception v3
6.2.4 DenseNet 121
6.3 Summary
6.4 Performance Evaluation Metrics
6.4.1 Confusion Matrix
6.4.2 Accuracy
6.4.3 Precision
6.4.4 Recall
6.4.5 F1 Score
6.4.6 Cross Entropy
6.5 Model Comparison
6.5.1 Validation Loss
6.5.2 Validation Accuracy
6.5.3 Accuracy Loss
6.6 Comparative Study
6.7 Training Mythology
6.8 Experiment Result
7. Conclusion

9
8. Future Scope
9. References
10. Appendices

ABSTRACT

10
 The process of choosing a college for higher education can be overwhelming and
challenging for students due to the plethora of options available. A college recommendation
system, powered by machine learning, can greatly assist students in making informed
decisions by providing personalized recommendations based on their preferences and
requirements.

This project aims to develop a college recommendation system that leverages machine
learning algorithms to analyze student data and recommend colleges that best match their
individual profiles. The system utilizes a variety of data, including academic performance,
extracurricular activities, interests, location preferences, and other relevant factors, to generate
accurate and customized recommendations.

The recommendation system employs a hybrid approach, combining both

collaborative filtering and content-based filtering techniques to enhance the accuracy and
diversity of recommendations. Collaborative filtering is utilized to identify similarities among
students based on their profiles and recommend colleges that similar students have shown
interest in. Content-based filtering is employed to analyze the features and characteristics of
colleges, such as location, accreditation, programs offered, faculty credentials, and student
reviews, to recommend colleges that align with the student's preferences and requirements.

To train and evaluate the recommendation system, a large dataset of student profiles
and college information is collected and preprocessed. Various machine learning algorithms,
including decision trees, support vector machines, and neural networks, are implemented and
compared to identify the most effective algorithm for the recommendation task. Feature
engineering techniques are also applied to enhance the system's performance and accuracy.

The developed college recommendation system offers several benefits to students,

educational institutions, and the broader educational ecosystem. For students, it provides
personalized recommendations that save time and effort in researching and evaluating
multiple colleges. It helps students make well-informed decisions based on their preferences
and increases the likelihood of finding a college that matches their needs and goals. For
educational institutions, the recommendation system can boost student enrollment and
retention rates by providing targeted recommendations that align with their programs and
offerings. It also enhances the overall student experience and satisfaction.

Furthermore, the recommendation system can contribute to the broader educational

ecosystem by providing valuable insights and data on student preferences, trends in college
choices, and areas of improvement for educational institutions. It can aid policymakers,
educational researchers, and administrators in making data-driven decisions to improve the
quality and accessibility of higher education.

11
 As a result, this project presents a college recommendation system that leverages
machine learning techniques to provide personalized and accurate recommendations to
students. The system utilizes a hybrid approach, combining collaborative filtering and content-
based filtering, and is trained and evaluated on a large dataset of student profiles and college
information. The system offers significant benefits to students, educational institutions, and
the broader educational ecosystem, and has the potential to positively impact the college
selection process for students worldwide. Further research and development can be
undertaken to continuously enhance the accuracy and effectiveness of the recommendation
system and extend its applicability to other domains in the educational landscape.

12
 CHAPTER 1

INTRODUCTION

1.1 Preface:

13
 In recent years, the landscape of higher education has undergone a significant
transformation, with a surge in demand for college and university admissions. Students today
are faced with an overwhelming number of choices when it comes to selecting a college or
university, making the decision-making process more complex than ever. In this era of
information overload, students need reliable and data-driven tools to assist them in making
informed decisions about their higher education options. This preface highlights the need for a
robust college recommendation system that leverages data preprocessing techniques to
provide accurate and personalized recommendations to students, ultimately aiding them in
making the right college choice.

The process of selecting a college or university has become increasingly complex and
challenging in recent years. With the expansion of higher education options and the
availability of vast amounts of information online, students are often overwhelmed by the
sheer volume of choices and the difficulty of evaluating various factors that influence their
decision. Factors such as location, campus facilities, rankings, courses offered, admission
requirements, and feedback from alumni are just a few of the many considerations that
students must weigh when making this important decision. However, the traditional methods
of gathering and analyzing this information, such as relying on rankings or recommendations
from peers, may not accurately reflect the individual needs and preferences of students,
leading to suboptimal choices and potential challenges in their academic journey.

In this context, the use of machine learning techniques and data preprocessing
approaches can provide a powerful solution to assist students in making informed decisions
about their college choices. By leveraging advanced data preprocessing techniques, such as
data cleansing, data transformation, feature engineering, and data integration, a college
recommendation system can effectively analyze and process vast amounts of data to generate
personalized recommendations tailored to the unique preferences and profiles of students.
Data preprocessing plays a crucial role in ensuring that the data used for generating
recommendations is of high quality, reliable, and relevant, and that it is processed in a manner
that reduces biases and promotes fairness.

The preface of this project sets the stage for the importance of data preprocessing in
the context of a college recommendation system. It highlights the challenges faced by students
in the college selection process and the limitations of traditional methods in providing
accurate and personalized recommendations. It also emphasizes the potential of machine
learning techniques and data preprocessing approaches to overcome these challenges and
provide students with a data-driven approach to college selection. The preface aims to create
awareness about the significance of data preprocessing and its relevance in the context of the
college recommendation system project, setting the stage for the subsequent sections that will
delve into the details of the project's objectives, methodology, and expected outcomes.

14
1.2 Brief Overview:

The college recommendation system project is a cutting-edge solution that aims to

utilize state-of-the-art data preprocessing techniques to analyze and process vast amounts of
data related to colleges. The project will leverage various data preprocessing algorithms to
cleanse, transform, and integrate data from multiple sources, including student profiles,
historical admission data, course offerings, campus facilities, location, rankings, and feedback
from alumni. The processed data will be used to generate personalized recommendations for
students based on their preferences, strengths, aspirations, and other relevant factors.

The data preprocessing phase of the project will involve several key techniques. Data
cleaning will be performed to identify and rectify inconsistencies, errors, and missing values
in the collected data. This may include handling outliers, dealing with duplicate or conflicting
data, and filling in missing values using appropriate imputation methods. Data transformation
will involve converting data into a consistent format, standardizing units, and normalizing
data for effective analysis. This may include converting categorical data into numerical
representations, scaling numerical data to a common range, and applying feature engineering
techniques to create new meaningful features from the raw data. Data integration will involve
combining data from multiple sources and creating a unified and comprehensive dataset for
analysis. This may include matching and merging data from different sources, resolving data
conflicts, and ensuring data consistency and integrity.

The college recommendation system will be designed to be user-friendly, with an

intuitive interface that allows students to input their preferences, view recommendations, and
compare different college options easily. The system will take into consideration various
factors, such as the student's academic performance, career goals, location preferences,
campus facilities, and other relevant criteria, to generate personalized recommendations that
align with the student's individual needs and aspirations. The recommendations will be
transparent, providing clear explanations of the factors that influenced the recommendations,
and allowing students to understand the basis for the system's suggestions.

Ethical considerations will also be a key focus of the project. Data privacy and security
will be carefully addressed, and appropriate measures will be implemented to protect sensitive
information and comply with relevant data protection regulations. Bias in data and
recommendations will be thoroughly addressed, and efforts will be made to ensure that the
system is fair, transparent, and unbiased. The project will strive to provide equal opportunities
and access to all students, regardless of their background or characteristics, and promote
diversity, equity, and inclusion in the college selection process.

The outcomes of this project are expected to have significant benefits for various
stakeholders. Students will have access to a reliable and data-driven tool that can provide them
with personalized recommendations, reducing the complexity and stress of the college
15
selection process. Educational institutions can benefit from increased visibility and improved
student engagement, as the system can provide insights into the preferences and requirements
of prospective students, helping them tailor their offerings to attract the right students.
Policymakers can also utilize the findings of this project to gain insights into the factors that
influence students' college choices and make informed decisions about education policies. The
project can also contribute to the field of data preprocessing by applying and advancing state-
of-the-art techniques in handling large and complex datasets, addressing data quality issues,
and integrating data from diverse sources.

The development of a robust college recommendation system that leverages advanced

data preprocessing techniques has the potential to revolutionize the way students make college
choices. The project aims to process vast amounts of data, cleanse and transform it, and
generate personalized and transparent recommendations for students. Ethical considerations,
such as data privacy, bias, and fairness, will be thoroughly addressed, and the system will be
designed to be user-friendly and accessible to all students. The outcomes of this project are
expected to have significant benefits for students, educational institutions, policymakers, and
the field of data preprocessing. With the potential to simplify the college selection process and
empower students to make informed decisions, this project has the potential to make a
meaningful impact on higher education.

16
 CHAPTER 2

LITERATURE REVIEW

The college selection process has become increasingly complex and challenging for
students due to the expansion of higher education options and the availability of vast amounts

17
of information online. In recent years, researchers have been exploring the use of machine
learning techniques to develop college recommendation systems that can assist students in
making informed decisions. Several studies have emphasized the importance of data
preprocessing in college recommendation systems. This involves cleaning, transforming, and
integrating data from multiple sources to improve system performance. Various techniques,
such as data normalization, feature selection, and data imputation, are used for effective data
preprocessing.

Machine learning techniques have been widely used in college recommendation

systems. These include collaborative filtering, content-based filtering, and hybrid approaches
that combine multiple methods. Collaborative filtering leverages user interactions and
similarity measures to provide recommendations, while content-based filtering uses item
features and user profiles. Hybrid approaches combine both collaborative and content-based
methods to overcome their limitations and improve recommendation accuracy.

The study by Sheetal Girase, Varsha Naik, and Debajyoti Mukhopadhyay proposes a
user-friendly college recommendation system that uses user profiling and matrix factorization.
User preferences are collected and combined with demographic information to create user
profiles. Matrix factorization, specifically Singular Value Decomposition (SVD), is applied to
uncover latent factors in the user-item preference matrix. The system generates
recommendations based on similarity between user profiles and college profiles. The study
finds that the proposed system is effective in providing personalized recommendations and
outperforms other methods in terms of accuracy and relevance. The contributions of the study
include a user-friendly system that addresses limitations of traditional methods and highlights
the importance of data preprocessing. The study is aligned with existing literature on college
recommendation systems that use similar approaches.

A college recommendation system proposed by Akshar Panchal, Kushal Gosar, Homil

Parmar, and Rohini Nair (2018) uses data mining and natural language processing techniques
to provide personalized recommendations to students. The system analyzes a dataset
containing information about colleges and user profiles to identify patterns and relationships
in the data. The study demonstrates the effectiveness of the proposed system in generating
relevant and accurate recommendations. It also highlights the importance of natural language
processing techniques in capturing meaningful information from textual data, such as college
descriptions and user preferences.

Other approaches in the field of college recommendation systems include collaborative

filtering and content-based filtering, which utilize similarity between users or colleges, and
features of colleges, respectively, to generate recommendations. Hybrid approaches
combining multiple techniques have also shown promising results. However, there are

18
limitations in the availability and accuracy of data, which should be considered in
implementing such systems.

The paper titled "The Institute Recommendation System Using Machine Learning" by
Asmita Orse, Nikhil Suryawanshi, Harsh Shrivastav, Pratik Bajpai, and Prof. Megha Patil,
published in 2022 with ISSN 2321-9653, presents a recommendation system for institutes
using machine learning techniques. The paper discusses the motivation behind the study, the
challenges associated with manual institute selection processes, and the potential benefits of
automated recommendation systems. The authors review existing research on
recommendation systems and machine learning, discussing different approaches and their
strengths and limitations. They emphasize the importance of data preprocessing and describe
the methodology of their proposed system, including data collection, preprocessing, feature
extraction, and model training. The authors evaluate the performance of their system using
metrics such as accuracy, precision, recall, and F1 score, and present their findings,
showcasing the effectiveness of their recommendation system. They compare their system
with other approaches, discuss strengths and limitations, and suggest potential areas of
improvement and future research directions.

The study by Subba Reddy and Govindarajulu proposes a college recommendation

system that uses student preferences and voting to generate recommendations. The system
collects user preferences through voting on college attributes and uses the data to generate
personalized recommendations. The empirical study shows that the system based on voting
data outperforms traditional collaborative filtering. The study highlights the benefits of
incorporating user preferences through voting, but acknowledges limitations such as potential
bias in the voting data and the need to consider other factors in the recommendation process.
Future research could explore ways to mitigate bias and incorporate additional factors for a
more comprehensive recommendation.

Udhayakumar and Harisai proposed a college recommendation system that uses data
mining techniques, specifically the collaborative filtering algorithm-2 (CF2), to generate
personalized recommendations for students. The study extensively analyzes CF2 and
compares it with other collaborative filtering algorithms. The findings suggest that CF2
outperforms other methods in terms of recommendation accuracy. The study also addresses
challenges such as data sparsity and scalability, proposing techniques like active learning and
cluster-based recommendation to overcome them.

The College Recommendation System proposed by Vinit Jain, Mohak Gupta, Jenish
Kevadia, Prof. Krishnanjalin Shinde (2017) proposes a college recommendation system that
uses content-based filtering approach, leveraging features of colleges such as location,
program offerings, accreditation, and faculty expertise. The system includes a user interface
for students to input their preferences, and personalized recommendations are generated based
19
on calculated similarity scores. The study provides insights into the development of content-
based filtering for college recommendation, considering specific preferences and requirements
of students in the college selection process.

"College Recommendation System using Machine Learning" by Jadhav Sayali

Pramod, Patil Durga Ananda, Kamble Sonali Arun, Shinde Sumit Sahebrao aims to develop a
college recommendation system using machine learning techniques. The authors propose a
hybrid approach that combines content-based filtering and collaborative filtering to generate
recommendations based on user preferences and college characteristics. Data preprocessing is
highlighted as crucial for reliable and consistent input data. The evaluation of the system using
metrics indicates high accuracy and relevance of recommendations, with the hybrid approach
outperforming individual techniques. Limitations include reliance on user data and not
considering other factors that may influence college choices. Further research could explore
incorporating additional factors for more comprehensive recommendations.

"Developing and Evaluating a University Recommender System" (2022) the study by

Mehdi Elahi, Alain Starke, Nabil El Loini, and Anna Alexander Lambrix, focuses on
developing and evaluating a university recommender system using collaborative filtering and
content-based filtering techniques. The study reviews existing literature on university
recommender systems and emphasizes the importance of personalized and relevant
recommendations for students. The study proposes a recommendation model that considers
user data, course data, and university data to generate recommendations based on user
preferences, course characteristics, and university features. The study shows that the hybrid
approach outperforms individual methods in terms of accuracy and relevance. However, the
study acknowledges limitations, such as reliance on user-generated data and potential data
sparsity.

The study titled "StudieMe: College Recommendation System" by Vidish Sharma,

Tarun Trehan, Rahul Chanana, and Suma Dawn proposes a college recommendation system
that uses machine learning techniques to generate personalized recommendations for students.
The study reviews existing literature on college recommendation systems, emphasizes the
need for effective personalized recommendations, and presents the architecture of the
StudieMe system, which includes data collection, data preprocessing, and recommendation
generation components. The study describes in detail the steps involved in each component
and evaluates the performance of the system using various metrics. The results show that
StudieMe performs well in terms of recommendation accuracy and relevance compared to
other systems. The study concludes by summarizing the findings, identifying limitations, and
suggesting future research directions. Overall, the study contributes to the field of college
recommendation systems by proposing a new system and highlighting its effectiveness in
providing personalized recommendations to students, while also acknowledging areas for

20
improvement and further research.

The study by Yara Zayed, Yasmeen Salman, and Ahmad Hasasneh (2022) proposes a
recommendation system that utilizes graduate student data to generate recommendations for
undergraduate programs at higher education institutions. The system leverages machine
learning techniques, specifically collaborative filtering, to provide personalized
recommendations based on the similarity between users' preferences and graduate students'
data. The study is unique in its use of graduate student data and builds on prior research in the
field of recommendation systems and higher education. The authors acknowledge the
limitations of using graduate student data and highlight the potential bias in the data. They
also suggest exploring other techniques such as content-based filtering and hybrid approaches
to enhance recommendation accuracy and relevance. The study contributes to the field by
proposing an innovative approach that can assist prospective undergraduate students in
making informed decisions about their academic and career paths.

The College Recommendation System is designed to assist students, parents, and other
entities in searching for top engineering colleges. By leveraging recommender system
techniques, this system aims to reduce the time and effort required for students to search and
compare colleges. The system utilizes data mining and machine learning algorithms to provide
personalized recommendations based on the data provided by users.

The system takes into consideration the information provided by students, such as their
preferences, location, and academic performance, and filters through the data to generate a list
of recommended colleges. This helps students in shortlisting colleges that align with their
preferences and requirements, making the college search process more efficient and effective.

The College Recommendation System is designed to be user-friendly and web-based,

making it easy to use and manage. Extensive testing has been conducted to ensure the
accuracy of the system, with results showing an accuracy rate of 89%. This indicates the
effectiveness of the system in providing relevant and reliable recommendations to users.
Overall, the designed system is a valuable tool for students, parents, and other users who are
searching for top engineering colleges. By leveraging data mining and machine learning
techniques, this system simplifies the college search process and provides accurate and
personalized recommendations, making it a helpful resource for decision-making.

21
 Chapter 7
CONCLUSION

The College Recommendation System is a user-friendly web program that has the
potential to significantly reduce the workload of students and simplify the process of selecting

22
the right college. Through extensive data preprocessing and the utilization of advanced
machine learning algorithms, the system offers personalized recommendations based on the
student's preferences, assisting them in making informed decisions and narrowing down their
choices of colleges.

The system serves as a valuable guide for learners, showcasing the potential of data
preprocessing and machine learning techniques in enhancing the decision-making process. It
aims to provide accurate and relevant recommendations, based on the student's academic,
career, and personal preferences, thus empowering students to make well-informed choices
about their higher education.

Furthermore, the College Recommendation System has a promising future scope for
improvement and expansion. By incorporating more advanced data preprocessing techniques,
exploring additional machine learning algorithms, and incorporating real-time data, the system
can further enhance the accuracy and relevance of the recommendations. Additionally, the
system can be expanded to include other streams and disciplines beyond engineering, making
it a comprehensive tool for a wider range of students.

The system represents a significant advancement in leveraging data preprocessing and

machine learning techniques to simplify the complex process of college selection for students.
It has demonstrated promising results in terms of accuracy, relevance, and usability during the
research project. By providing a user-friendly web program that offers personalized
recommendations based on the student's preferences, the system aims to alleviate the burden
and confusion associated with choosing the right college for higher education.

The system has the potential to create a positive impact on students' lives by guiding
them towards colleges that align with their academic, career, and personal aspirations. Its
intuitive user interface, interactive features, and comprehensive database of colleges make it a
valuable tool for students in their college search journey. The system has the potential to
contribute to improving college admission outcomes by assisting students in selecting colleges
that are a good fit for their interests, goals, and preferences.

In conclusion, the College Recommendation System has the potential to revolutionize

the college selection process by reducing students' workload, simplifying decision-making,
and providing personalized recommendations. Its utilization of data preprocessing and
machine learning methods, its value as a guide for learners, and its future scope for
improvement make it a valuable contribution to the field of data preprocessing and machine
learning in the context of college admissions.

23
 Chapter 8
FUTURE SCOPE

The future scope of the college recommendation system holds immense potential for
advancements in data preprocessing and machine learning techniques. By incorporating multi-
dimensional recommendations, adaptive and contextual recommendations, social and peer
recommendations, predictive analytics, enhanced personalization, integration with virtual and

24
augmented reality, continuous monitoring and feedback, and expansion to global
recommendations, the system can evolve into a powerful tool for assisting students in making
informed decisions about their college choices.

Multi-dimensional Recommendations: In the future, the college recommendation system

could incorporate multi-dimensional recommendations, taking into account not only the
academic preferences of students but also other factors such as extracurricular activities,
sports, cultural events, and community involvement. This could help students find colleges
that align with their overall interests and personality, providing a more holistic approach to
college selection.

Adaptive and Contextual Recommendations: The system could further evolve to provide
adaptive and contextual recommendations based on the changing needs and preferences of
students. For example, it could adapt recommendations based on the latest trends in the job
market, emerging technologies, or global events, ensuring that students receive up-to-date and
relevant recommendations for their future career prospects.

Social and Peer Recommendations: The system could incorporate social and peer
recommendations, leveraging social networks or student communities to provide
recommendations based on the experiences and feedback of current students or alumni. This
could provide valuable insights and perspectives from individuals who have firsthand
knowledge of the colleges, helping students make more informed decisions.

Predictive Analytics: With advancements in machine learning and predictive analytics, the
system could analyze historical data on student admissions, academic performance, and career
outcomes to predict the likelihood of a student's success in a particular college or program.
This could assist students in making data-driven decisions and selecting colleges that align
with their long-term career goals.

Enhanced Personalization: The system could further enhance personalization by

incorporating additional data sources, such as student profiles, learning styles, and cognitive
abilities, to provide tailored recommendations that meet the unique needs and preferences of
individual students. This could result in a more personalized and engaging experience for
students, increasing the system's effectiveness in helping them find the right college fit.

Integration with Virtual and Augmented Reality: As virtual and augmented reality
technologies continue to advance, the system could potentially integrate with these
technologies to provide virtual tours, simulations, or immersive experiences of college
campuses. This could allow students to virtually explore colleges and get a better sense of the
campus environment, facilities, and culture, helping them make more informed decisions.

25
Continuous Monitoring and Feedback: The system could incorporate continuous
monitoring and feedback mechanisms to collect data on student outcomes, satisfaction, and
feedback on recommended colleges. This could help in evaluating the effectiveness of the
system and making regular updates and improvements to ensure its accuracy and relevance for
students.

Expansion to Global Recommendations: The system could expand its scope beyond local
colleges and include recommendations for international colleges and universities. This could
be particularly beneficial for students interested in pursuing higher education abroad,
providing them with relevant recommendations based on factors such as location, reputation,
accreditation, and cost.
These advancements have the potential to greatly enhance the accuracy, relevance, and
effectiveness of the system, ultimately benefiting students by providing them with tailored
recommendations that align with their academic, career, and personal aspirations. As
technology continues to evolve, the college recommendation system can play a pivotal role in
helping students navigate the complex landscape of college selection and maximize their
chances of finding the right college fit for their future success.

26
 Chapter 9
REFERENCES

27
1] Sheetal Girase, Varsha Naik, Debajyoti Mukhopadhyay, “A User-friendly College
Recommending System using User-profiling and Matrix Factorization Technique”,
DOI: 10.1109/CCAA.2017.8229779, 2017.
2] Akshar Panchal, Kushal Gosar, Homil Parmar, Rohini Nair, “College
Recommendation System using Data Mining and Natural Language Processing”, 2018.
3] Asmita Orse, Nikhil Suryawanshi, Harsh Shrivastav, Pratik Bajpai, Prof. Megha Patil,
“Institute Recommendation System Using ML”, ISSN: 2321-9653, 2022.
4] Mr. Y. Subba Reddy and Prof. P. Govindarajulu, “College Recommender system using
student’ preferences/voting: A system development with empirical study”, 2018.
5] Udhayakumar S, Harisai V, “College Recommendation System For Students Using
Datamining With Collaborative Filtering Algorithm–2”, ISSN: 2582-5208 , 2021.
6] Vinit Jain, Mohak Gupta, Jenish Kevadia, Prof. Krishnanjalin Shinde, “College
Recommendation System”, ISSN: 2278-0181, 2017.
7] Jadhav Sayali Pramod , Patil Durga Ananda , Kamble Sonali Arun, Shinde Sumit
Sahebrao, “College Recommendation System using Machine Learning”, -ISSN: 2395-
0056, Mar 2020.
8] “Developing and Evaluating a University Recommender System”, Mehdi Elahi, Alain
Starke, Nabil El loini, Anna Alexander Lambrix, 2022.
9] Vidish Sharma, Tarun Trehan, Rahul Chanana, Suma Dawn, “StudieMe: College
Recommendation System”, DOI: 10.1109/RDCAPE47089.2019.8979030, 2019.
10] “A Recommendation System for Selecting the Appropriate Undergraduate Program at
Higher Education Institutions Using Graduate Student Data”, Yara Zayed, Yasmeen
Salman, Ahmad Hasasneh, https://doi.org/10.3390/app122412525, 2022.

28
 Chapter 10
APPENDICES

29
30
31
 Instructions for Project report writing
Sequence of pages –
Part A
1 Cover page (Without copyright)
2 Inner first page (With copyright) ©G H Raisoni College of Engineering, Nagpur 2021
3 Declaration & Certificate – Color print with RGI logo as watermark(do not stretch)
4 Certificate from industry if any on industry letterhead describing duration and project work done,
achievement & stipend/placement details.
5 Acknowledgement
Note - Cover page & Inner page will be same except inner page will have copyright instruction
Part B (This part will have page numbers in roman i,e. i,ii,iii………..)
5 Abstract (1-2 pages on Complete Project) --- Roman page nos. (Start)
6 List of Figures
7 List of Tables
8 List of Symbols
9 List of Publications/Project competitions/Patents filed if any. Roman page nos. (End)
Part C (This part will have page numbers as 1,2,3…………..)
10 Index
11 Chapters(Numerals/page no. From 2 page onwards of every chapter and separator between chapters)
nd

Tentative Chapter Scheme

1. Introduction
2. Literature Review (10-14 pages minimum)
3. Methodology
4. Data Collection / Tools / Platform used
5. Design / Implementation / modelling
6. Testing & Summary of Results
7. Conclusion (should be in bulleted form)
8. Future Scope (should be in bulleted form)
9. Internship hands on practice / work done at industry /case studies undertaken by all students of group
10. References (IEEE/ASME/ASCE style in chronological order). Refer IEEE/ASME/ASCE papers no other
format will be allowed.
Appendices
A. List of images/photos of project, group photo with guide/co-guide (if any)
B. List of image/photos in industry along with industry supervisors
C. Copy of certificate of winning/participation in project competition/best innovation awards
D. Paper published if any
E. Soft copy of entire project report & PPT in CD& Video of project of 1 Min.

Paper size A 4, Margin: Top 1, Bottom 1.25, Left 1.5, Right 0.8,
Page Nos. Bottom center,Spacing 1.5
Chapter & chapter title 16 Bold Upper case
Section & section title 14 Bold Title case
Sub-section & Sub-section title 12 Bold Title case
Title of figures, tables 12 Bold
Other written matter 12 Normal

32