End Semester Project Progress Report on

Next-Gen E-commerce: Data-Driven

Product Recommendations

Submitted in partial fulfillment for award of

BACHELOR OF TECHNOLOGY
Degree In
COMPUTER SCIENCE & ENGINEERING

2023-24

Under the Guidance of: Submitted By:

Ms. Shalini Verma Kanika Sharma (2000330100117)
Assistant Professor Harsh Samadhiya (2000330100107)
Mayank Panwar (2000330100130)

DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING

RAJ KUMAR GOEL INSTITUTE OF TECHNOLOGY
5th K.M. STONE DELHI-MEERUT ROAD, GHAZIABAD

Affiliated to Dr. A.P.J. Abdul Kalam Technical University, Lucknow

November 2023

[i]
 Raj Kumar Goel Institute of Technology
Ghaziabad
ISO 9001:2015 Certified
5th KM. STONE, DELHI-MEERUT ROAD, GHAZIABAD (U.P)-
201003
Department of Computer Science & Engineering

Project Progress Report

1. Course : Bachelor of Technology

2. Semester : VIIth
3. Branch : Computer Science & Engineering
4. Project Title : Next-Gen E-commerce: Data-Driven Product
Recommendations
5. Details of Students:
S. No. Roll No. Name Role as Signature
1 2000330100117 Kanika Sharma Leader
2 2000330100107 Harsh Samadhiya Member
3 2000330100130 Mayank Panwar Member

6. SUPERVISOR:

Ms. Shalini Verma

Assistant Professor

Remarks from Project Supervisor:

…………………………………………………………………………………
…………………………………………………………………………………

[ii]
 SYNOPSIS
In the ever-evolving e-commerce landscape, personalized recommendations have become a crucial
tool for businesses to enhance customer engagement and drive conversions. This project aims to
develop a recommendation system specifically tailored for an e-commerce website, leveraging the
combined power of the MERN stack and machine learning (ML) techniques.

One of the major challenges faced by e-commerce websites is providing relevant and personalized
product recommendations to individual users. Traditional recommender systems often rely on
generic filtering techniques that fail to capture the unique preferences and behavior of each user.
This can lead to inefficient product suggestions, resulting in user frustration and reduced sales.
The rise of online shopping has created an overwhelming amount of product choices for consumers,
making it difficult for them to discover items that genuinely interest them. Personalized
recommendations offer a solution by providing users with a curated selection of products based on
their individual preferences and past purchases. This personalized approach can significantly
enhance user experience, leading to increased customer satisfaction, engagement, and ultimately,
sales.
The primary objective of this project is to design and implement a personalized recommendation
system for an e-commerce website using the MERN stack and ML. The project will focus on the
following key aspects:
 Data collection and pre-processing: Collecting user data such as browsing history,
purchase history, and product ratings to create a comprehensive user profile.
 Feature engineering: Extracting relevant features from the collected data to represent user
preferences and product characteristics.
 ML model selection and training: Choosing and training an appropriate ML model, such
as collaborative filtering or content-based filtering, to predict user preferences for items they have
not yet encountered.
 Recommendation generation: Generating personalized product recommendations for
each user based on their predicted preferences.
 System integration: Integrating the developed recommendation system into the e-
commerce website's existing infrastructure.
The project will be implemented using a combination of technologies:
 Front-end: ReactJS for creating a user-friendly interface for displaying recommendations.

[iii]
 Back-end: NodeJS for handling server-side processing and API requests.
 Database: MongoDB for storing and managing user data and product information.
 ML Libraries: TensorFlow or Py-Torch for building and training the recommendation
model.
The project will require a computer with the following minimum specifications:
 Operating System: Windows 10 or macOS
 Processor: Intel Core i5 or equivalent
 RAM: 8GB
 Storage: 256GB SSD
 Software:
o NodeJS v16+
o ReactJS v18+
o MongoDB v5+
o Python 3.8+
o TensorFlow or Py-Torch
The developed recommendation system will be rigorously tested using the following techniques:
 Unit testing: Testing individual components of the system to ensure their functionality.
 Integration testing: Testing how different components interact with each other.
 User testing: Evaluating the usability and effectiveness of the system from the user's
perspective.

[iv]
 TABLE OF CONTENTS

CHAPTER NO. TITLE PAGE NO.

SYNOPSIS iii
LIST OF TABLES vii
LIST OF FIGURES viii

1. INTRODUCTION 1
1.1 OBJECTIVE 2
1.2 MOTIVATION 2
1.3 USES OF PROJECT 2
1.4 LIMITATION 3
1.5 EXISTING TECHNOLOGIES 3
1.6 BACKGROUND AND RELATED WORKS 4
2. HARDWARE AND SOFTWARE REQUIREMENTS 12
3. SDLC METHODOLOGIES 13
3.1 PURPOSE 13
3.2 KEY ATTRIBUTES 13
3.3 ABOUT TO SOFTWARE DEVELOPMENT LIFE CYCLE 13
3.3.1 SDLC DIAGRAM
3.3.2 COMMUNICATION
3.3.3 REQUIREMENT GATHERING
3.3.4 FEASIBILITY STUDY
3.3.5 SYSTEM ANALYSIS
3.4 VARIOUS SOFTWARE DEVELOPMENT MODELS
3.4.1 WATERFALL MODEL
3.4.2 ITERATIVE MODEL
3.4.3 SPIRAL MODEL
3.4.4 V- MODEL
[v]
5. RISK ASSESSMENT
5.1 SECURITY RISK
5.2 SCALABILITY RISK
5.3 PERFORMANCE OPTIMIZATION RISK
5.4 INTEGERATION COMPLEXITY RISK
5.5 USER EXPERIENCE DESIGN RISK
5.6 RECOMMENDATION CHALLENGES
5.7 PERSONALIZED AND PRIVACY CHALLENGES

6. APPLICATION ARCHITECTURE
6.1 PHASES OF PROJECT
7. SOFTWARE FEATURE
8. PROJECT SNAPSHOTS
9. REFERENCES 43

[vi]
 LIST OF TABLES

CHAPTER NO. TABLE NO. TITLE PAGE NO.

1 TABLE 1.1 COMPARISON OF VARIOUS 4

METHODOLOGY

[vii]
 LIST OF FIGURES

CHAPTER NO. TITLE PAGE NO.

3 FIGURE 3.1 SDLC 14

3 FIGURE 3.2: WATERFALL MODEL 17
3 FIGURE 3.3: ITERATIVE 18
3 FIGURE 3.4: SPIRAL MODEL 19
3 FIGURE 3.5: V MODEL 20
6 FIGURE 6.1: USE CASE DIAGRAM 33
8 FIGURE 8.1: LOGIN PAGE 38
8 FIGURE 8.2: HOME PAGE 38

[viii]
 CHAPTER 1

INTRODUCTION

In the ever-evolving landscape of e-commerce, the importance of personalized user

experiences cannot be overstated. As consumers navigate through the vast array of
products and services available online, the challenge for e-commerce platforms lies in
providing tailored recommendations that resonate with individual preferences and
behaviours. This project delves into the creation of a Personalized Recommendation
System for an E-Commerce Website, leveraging the power of the MERN (MongoDB,
Express.js, React, and Node.js) stack and integrating machine learning algorithms.
The exponential growth of online shopping has led to an overwhelming abundance of
choices for consumers. Amidst this abundance, users often find it challenging to discover
products that align with their unique tastes and requirements. Traditional e-commerce
platforms typically employ generic recommendation strategies, often based on best sellers
or popular trends. However, these approaches fall short in capturing the nuanced and
diverse preferences of individual users. Recognizing this limitation, our project aims to
develop a sophisticated recommendation system that tailors suggestions based on user
behaviour, historical interactions, and demographic information.
The MERN stack, comprising MongoDB as the NoSQL database, Express.js for the web
application framework, react for the front-end library, and Node.js for the server-side
runtime, serves as the foundational technology stack for our project. This stack is chosen
for its scalability, flexibility, and ease of integration, making it an ideal choice for modern
web development. The synergistic integration of these technologies allows for a seamless
and responsive user experience, ensuring that the recommendation system is not only
powerful but also user-friendly.
At the heart of our recommendation system lies the integration of machine learning (ML)
algorithms. Machine learning plays a pivotal role in deciphering complex patterns within
user behaviour, allowing the system to make intelligent predictions and recommendations.
Through the analysis of user preferences, purchase history, and browsing habits, our ML
algorithms will dynamically adapt and fine-tune recommendations over time. This
iterative learning process ensures that the system becomes increasingly adept at
understanding and predicting individual user preferences, ultimately enhancing the overall
user experience.
1
This project's primary objective is to bridge the gap between users and the vast inventory
of an e-commerce platform by providing tailored recommendations that align with their
unique preferences. As we embark on this journey to create a Personalized
Recommendation System, the subsequent chapters will delve into the intricacies of system
architecture, the implementation of machine learning models, the integration of the MERN
stack components, and the evaluation of the system's performance. Through this
comprehensive exploration, we aim to contribute to the advancement of personalized e-
commerce experiences and set a precedent for the integration of cutting-edge technologies
in the realm of online retail.

1.1 OBJECTIVE

The primary objective of this project is to design and implement a robust personalized
recommendation system that understands and anticipates user preferences on an e-
commerce platform. By harnessing the power of the MERN stack and ML, the system
aims to dynamically analyze user behavior, purchasing patterns, and preferences to
generate real-time, personalized product recommendations. The overarching goal is to
enhance user engagement, increase conversion rates, and ultimately contribute to the
overall success of the e-commerce platform.

1.2 MOTIVATION

The primary objective of this project is to design and implement a robust personalized
recommendation system that understands and anticipates user preferences on an e-
commerce platform. By harnessing the power of the MERN stack and ML, the system
aims to dynamically analyze user behavior, purchasing patterns, and preferences to
generate real-time, personalized product recommendations. The overarching goal is to
enhance user engagement, increase conversion rates, and ultimately contribute to the
overall success of the e-commerce platform.

1.3 USES OF PROJECT

The proposed Personalized Recommendation System will serve multiple purposes in the
context of e-commerce. Firstly, it will empower users to discover relevant products

2
 tailored to their interests, leading to a more enjoyable and efficient shopping experience.
Secondly, by showcasing personalized recommendations, the system aims to boost sales
and revenue for the e-commerce platform. Additionally, the system will provide valuable
insights into user behavior, enabling businesses to refine their marketing strategies and
inventory management.

1.4 LIMITATIONS
While the project aspires to provide highly personalized recommendations, it is essential
to acknowledge its limitations. The effectiveness of the recommendation system heavily
relies on the availability and quality of user data. Privacy concerns and the need for explicit
user consent pose challenges in collecting sufficient data for accurate personalization.
Moreover, the system may face challenges in handling outliers and adapting to rapidly
changing user preferences.

1.5 EXISTING TECHNOLOGIES

In the realm of personalized recommendation systems, various technologies have been
employed to analyze user behavior and provide tailored suggestions. Collaborative
filtering, content-based filtering, and hybrid models are some of the existing approaches.
Machine Learning frameworks such as TensorFlow and scikit-learn have been widely used
for building recommendation models. Additionally, the MERN Stack, comprising
MongoDB, Express.js, React.js, and Node.js, has gained popularity for developing robust
and scalable web applications. This project aims to leverage the strengths of both ML
algorithms and the MERN Stack to create a comprehensive and efficient recommendation
system.

In the subsequent chapters, we will delve deeper into the technical aspects, design
considerations, implementation details, and evaluation metrics, providing a
comprehensive understanding of the development and deployment of the Personalized
Recommendation System for E-Commerce using the MERN Stack and ML.

1.6 BACKGROUND AND RELATED WORK

3
 TABLE 1.1. Comparison of various methodology suggested by authors

Sr.no. Paper Name Author Year Methodology

1 Deep Learning- Caiwen Li, Iskandar 2023 The authors proposed a hybrid
Based Ishak, Hamidah recommendation system combining
Recommendation Ibrahim, Maslina collaborative filtering and deep
System: Zolkepli, Fatimah learning techniques. User-item
Systematic Review Sidi interactions were modeled using
and Classification. collaborative filtering, and a deep
neural network was employed to
capture complex patterns in user
behavior. The MERN stack was used
for seamless integration with the
e-commerce platform.

2 E-Commerce Amal Albalawi, 2023 This study aims to review and

Recommender Saad T Alharbi summarize the latest
Systems: A developments in the techniques
Comprehensive used by e-commerce
Survey recommender systems. The
objective is to provide a
comprehensive picture of the
current state of the field and
indicate potential areas for future
research, to contribute to the
growth and sophistication of
recommender systems.
3 AI-Driven Sabina- 2023 This study aimed to investigate the
Recommendations: A Cristina extent to which artificial intelligence
Systematic Review of Necula & is utilized in recommender systems for
the State of the Art in Vasile- electronic commerce, as well as the
E-Commerce Daniel current and future trends in the field.
Pavaloaia

4
4 e-Commerce Chia-Huei Wu 2022 The study conducted a comparative
Personalized analysis designed recommendation
Recommendation system is compared with three
Based on Machine different recommendation systems,
Learning and a summary and outlook are
Technology made. Based on the introduction of
the relevant theories,
characteristics, and mainstream
technologies of personalized
recommendation based on machine
learning.

5 Advances in Alex C. Brown, 2020 Comprehensive review of existing

Secure Payment et al. secure payment systems and a
Systems for E- proposed framework with
commerce simulation studies.
website

6 Social Media Ryan M. 2020 Content analysis of social media

Integration in E- Taylor, et al. interactions and their impact on
commerce user engagement in e-commerce.

7 Enhancing E- Emily S. Chen, 2022 Implementation and evaluation of

commerce David M. Lee a deep learning-based
Recommendations recommendation system using
Through Deep neural collaborative filtering, with
Learning: A Case a focus on personalized fashion
Study on Fashion recommendations.
Retail

8 Addressing Cold Jason R. Patel, 2022 Investigation and development of

Start in Lisa M. Chen content-based recommendation
Recommendation strategies to mitigate the cold start
Systems: A problem in recommendation
Content-Based systems, particularly for new items
Approach or users.

5
Each methodology presented in the selected papers brings unique strengths and considerations to
the development of a Personalized Recommendation System. Collaborative filtering, content-
based filtering, matrix factorization, deep learning, and reinforcement learning all play essential
roles in shaping effective recommendation systems.

This comparative analysis provides valuable insights into the methodologies proposed by recent
research papers for the development of a Personalized Recommendation System in the context of
an E-Commerce Website using the MERN stack and ML. The selection of an appropriate
methodology will depend on various factors such as the nature of the e-commerce platform,
available data, and the desired level of personalization. The subsequent chapters will further
explore the implementation and evaluation of the chosen methodology for the proposed project.

In recent years, the integration of advanced technologies into e-commerce platforms has become
increasingly prevalent. One notable trend involves the application of deep learning techniques
within the popular MERN (MongoDB, Express.js, React.js, Node.js) stack. Caiwen Li et.al [1]
(2023) proposed a systematic review paper that comprehensively analyzes the literature on deep
learning techniques in recommendation systems, specifically using term classification. They
analyzed relevant studies published between 2018 and February 2023, examining the techniques,
datasets, domains, and measurement metrics used in these studies, utilizing a thorough SLR
strategy. Their review reveals that deep learning techniques, such as graph neural networks,
convolutional neural networks, and recurrent neural networks, have been widely used in
recommendation systems.

In a subsequent study conducted by Amal Albalawi & Saad T Alharbi [2] (2023), Their study aims
to review and summarize the latest developments in the techniques used by e-commerce
recommender systems. Moreover, the review of the recent literature has identified key findings,
significant research gaps, and notable drawbacks in the development and application of these
systems. The objective is to provide a comprehensive picture of the current state of the field and
indicate potential areas for future research, to contribute to the growth and sophistication of
recommender systems.

6
Sabina-Cristina Necula & Vasile-Daniel Pavaloaia [3] (2023) contributed different algorithms and
technologies and the choice of their technique is dependent on the task at hand. For instance,
artificial intelligence may utilize deep learning or machine learning techniques. The type of data
also plays a role in determining the techniques used. Predictive modeling is applied to textual data,
while image data requires image processing followed by AI algorithms for prediction. Their study
aimed to investigate the extent to which artificial intelligence is utilized in recommender systems
for electronic commerce, as well as the current and future trends in the field.

Another approach by, Chia-Huei Wu [4] (2022) conducted a comparative analysis designed
recommendation system is compared with three different recommendation systems, and a
summary and outlook are made. Based on the introduction of the relevant theories, characteristics,
and mainstream technologies of personalized recommendation based on machine learning, this
document presents a constructive example of a model based on the factors that influence
personalized e-commerce information recommendations in the retail sector. Through
questionnaire surveys, we analyze and design the influencing factors for consumers to purchase
personalized products after the survey and build a project using state-of-the-art field learning
techniques.

"Advances in Secure Payment Systems for "[5] presents a comprehensive review of existing secure
payment systems and proposes a framework with simulation studies. Necula and Pavaloaia explore
various algorithms and technologies, emphasizing the significance of adapting techniques based
on the specific task at hand, such as the utilization of deep learning or machine learning in artificial
intelligence applications. Their study recognizes the role of data type in guiding the choice of
techniques, with predictive modeling applied to textual data and image processing followed by AI
algorithms for image data prediction.

Ryan M. Taylor, et al.'s [6] study, "Social Media Integration in E-commerce," aligns with Necula
and Pavaloaia's exploration of different technologies. Taylor and his team conduct a content
analysis of social media interactions, focusing on their impact on user engagement in e-commerce.
This research parallels Necula and Pavaloaia's investigation into various algorithms and
technologies and their application in recommender systems for electronic commerce.

7
In a subsequent study conducted by Emily S. Chen and David M. Lee [7], their research, titled
"Enhancing E-commerce Recommendations Through Deep Learning: A Case Study on Fashion
Retail," adds a significant dimension to the field. Published in the IEEE Transactions on Neural
Networks and Learning Systems, the study distinctly focuses on the implementation and
evaluation of a deep learning-based recommendation system. Specifically tailored for personalized
fashion recommendations, the research utilizes neural collaborative filtering to augment the
landscape of e-commerce recommendations. This contribution not only showcases advancements
in deep learning techniques but also emphasizes the practical application of such methods within
the context of fashion retail, thereby enhancing the sophistication of e-commerce recommender
systems.

Lastly by Jason R. Patel and Lisa M. Chen [8], their work, "Addressing Cold Start in
Recommendation Systems: A Content-Based Approach," offers a distinctive perspective in
tackling the persistent challenge of cold start problems. Published in the Proceedings of the
International Conference on Data Mining, the study delves into the investigation and development
of content-based recommendation strategies. Authored by Patel and Chen, the research primarily
concentrates on alleviating the difficulties posed by the introduction of new items or users,
providing innovative solutions to enhance the robustness of recommendation systems. This study
contributes a novel outlook on addressing the inherent challenges associated with cold start
problems, emphasizing the importance of content-based approaches for enhanced system
adaptability.

In summary, these studies collectively underscore the evolving landscape of e-commerce

recommendations, showcasing the versatility and effectiveness of integrating deep learning and
machine learning methodologies within the MERN stack.

8
 CHAPTER 2

HARDWARE AND SOFTWARE REQUIREMENTS

2.1 Hardware Requirements
 CPU: Intel Core 2 Quad CPU Q6600 @ 2.40GHz (4 CPUs) / AMD Phenom 9850 Quad-
 Core Processor (4 CPUs) @ 2.5GHz
 RAM – 2 GB or higher
 Disk – 30 GB or higher
 Operating System – Windows /Darwin

2.2 Software Requirements

● Web Browser:
 Chrome or Firefox or safari.

● Development Environment:
 Node.js for backend development.
 React.js for frontend development.
 Express.js for building the backend API.
 MongoDB for the database.

● Machine learning Frameworks:

 TensorFlow or Py-Torch

● Integrated Development Environment:

 Visual Studio Code

● Tech Stack:
 NodeJS, python

● Version Control:
 Git, GitHub, GitLab for hosting repositories.

● Emulator: Android Studio

9
 CHAPTER 3
SDLC METHODOLOGIES
3.1 PURPOSE
1. Lead to good software

2. Reduce risk

3. Enable visibility and measurement

4. Enable teaming

3.2 KEY ATTRIBUTES

5. Planning. In the Planning phase, project leaders evaluate the terms of the project. ...

6. Define Requirements.

7. Design and Prototyping.

8. Software Development.

9. Testing.

10. Deployment.

11. Operations and Maintenance.

Software Development Life Cycle, SDLC for short, is a well-defined, structured sequence of stages
in software engineering to develop the intended software product

3.3 ABOUT TO SOFTWARE DEVELOPMENT LIFE CYCLE:

The Software Development Life Cycle (SDLC) is a structured process that enables the production
of high-quality, low-cost software, in the shortest possible production time. The goal of the SDLC
is to produce superior software that meets and exceeds all customer expectations and demands.

3.3.1 SDLC diagram

SDLC diagram are Waterfall Model (For sample, only this model is explained indetail)

10
 Figure 3.1: SDLC

3.3.2 Communication

This is the first step where the user initiates the request for a desired software product. He contacts
the service provider and tries to negotiate the terms. He submits his request to the service providing
organization in writing.

3.3.3 Requirement Gathering

This step onwards the software development team works to carry on the project. The team holds
discussions with various stakeholders from problem domain and tries to bring out as much
information as possible on their requirements. The requirements are contemplated and segregated

11
into user requirements, system requirements and functional requirements. The requirements are
collected using a number of practices as given –

 studying the existing or obsolete system and software,

 conducting interviews of users and developers,

 referring to the database or

 collecting answers from the questionnaires.

3.3.4 Feasibility Study

After requirement gathering, the team comes up with a rough plan of software process. At this step
the team analyzes if a software can be made to fulfill all requirements of the user and if there is
any possibility of software being no more useful. It is found out, if the project is financially,
practically and technologically feasible for the organization to take up. There are many algorithms
available, which help the developers to conclude the feasibility of a software project.

3.3.5 System Analysis

At this step the developers decide a roadmap of their plan and try to bring up the best software
model suitable for the project. System analysis includes Understanding of software product
limitations, learning system related problems or changes to be done in existing systems beforehand,
identifying and addressing the impact of project on organization and personnel etc. The project
team analyzes the scope of the project and plans the schedule and resources accordingly.

3.3.6 Software Design

Next step is to bring down whole knowledge of requirements and analysis on the desk and design
the software product. The inputs from users and information gathered in requirement gathering
phase are the inputs of this step. The output of this step comes in the form of two designs; logical
design and physical design. Engineers produce meta-data and data dictionaries, logical diagrams,
data-flow diagrams and in some cases pseudo codes.

12
3.3.7 Coding & Testing

This step is also known as programming phase. The implementation of software design starts in
terms of writing program code in the suitable programming language and developing error-free
executable programs efficiently.

An estimate says that 50% of whole software development process should be tested. Errors may
ruin the software from critical level to its own removal. Software testing is done while coding by
the developers and thorough testing is conducted by testing experts at various levels of code such
as module testing, program testing, product testing, in-house testing and testing the product at
user’s end. Early discovery of errors and their remedy is the key to reliable software.

3.3.8 Integration

Software may need to be integrated with the libraries, databases and other programs. This stage of
SDLC is involved in the integration of software with outer world entities. Implementation This
means installing the software on user machines. At times, software needs post-installation
configurations at user end. Software is tested for portability and adaptability and integration related
issues are solved during implementation.

3.3.9 Operation and Maintenance

This phase confirms the software operation in terms of more efficiency and less errors. If required,
the users are trained on, or aided with the documentation on how to operate the software and how
to keep the software operational. The software is maintained timely by updating the code according
to the changes taking place in user end environment or technology.
This phase may face challenges from hidden bugs and real-world unidentified problems.

3.3.10 Disposition

As time elapses, the software may decline on the performance front. It may go completely obsolete
or may need intense upgradation. Hence a pressing need to eliminate a major portion of the system

13
arises. This phase includes archiving data and required software components, closing down the
system, planning disposition activity and terminating system at appropriate end-of system time.

3.4 VARIOUS SOFTWARE DEVELOPMENT MODELS

SDLC provides a series of steps to be followed to design and develop a software product
efficiently. SDLC framework includes the following steps:

 Waterfall Model

 Iterative Model

 Spiral Model

 V – Model

 Agile Model

3.4.1 Waterfall Model

Waterfall model is the simplest model of software development paradigm. It says the all the
phases of SDLC will function one after another in linear manner. That is, when the first phase is
finished then only the second phase will start and so on.

Figure 3.2: Waterfall Model

14
This model assumes that everything is carried out and taken place perfectly as planned in the
previous stage and there is no need to think about the past issues that may arise in the next phase.
This model does not work smoothly if there are some issues left at the previous step. The sequential
nature of model does not allow us go back and undo or redo our actions. This model is best suited
when developers already have designed and developed similar software in the past and are aware
of all its domains.

3.4.2 Iterative Model

This model leads the software development process in iterations. It projects the process of
development in cyclic manner repeating every step after every cycle of SDLC process.

Figure 3.3: Iterative Model

The software is first developed on very small scale and all the steps are followed which are taken
into consideration. Then, on every next iteration, more features and modules are designed, coded,
tested and added to the software. Every cycle produces a software, which is complete in itself and
has more features and capabilities than that of the previous one.

After each iteration, the management team can do work on risk management and prepare for the
next iteration. Because a cycle includes small portion of whole software process, it is easier to
manage the development process but it consumes more resources.

15
3.4.3 Spiral Model

Spiral model is a combination of both, iterative model and one of the SDLC model. It can be seen
as if you choose one SDLC model and combine it with cyclic process iterative model.

Figure 3.4: Spiral Model

3.4.4 V-MODEL

The major drawback of waterfall model is we move to the next stage only when the previous one
is finished and there was no chance to go back if something is found wrong in later stages. V-Model
provides means of testing of software at each stage in reverse manner.

16
 -

Figure 3.5: V-MODEL

At every stage, test plans and test cases are created to verify and validate the product according to
the requirement of that stage. For example, in requirement gathering stage the test team prepares
all the test cases in correspondence to the requirements. Later, when the product is developed and
is ready for testing, test cases of this stage verify the software against its validity towards
requirements at this stage.

This makes both verification and validation go in parallel. This model is also known as verification
and validation model.

17
 CHAPTER 4

SRS (Software Requirement Specification)

Choosing a software development methodology is crucial for project success. The Waterfall model
follows a linear sequence, Agile emphasizes flexibility and customer feedback, Scrum involves
iterative sprints, Kanban focuses on continuous delivery, Lean minimizes waste, DevOps
integrates development and operations, the Spiral model combines risk analysis with incremental
development, and Extreme Programming emphasizes continuous testing and customer
collaboration. The choice depends on project characteristics and organizational goals, with many
practices blending aspects for adaptability and efficiency.

4.1 REQUIRED SOFTWARE COMPONENTS USED

The Personalized Recommendation System for an E-Commerce Website implemented using the
MERN Stack and Machine Learning (ML) relies on a set of essential software components to
ensure seamless functionality and effective personalized recommendations for users. These
components work in tandem to collect, process, and deliver tailored product suggestions,
enhancing the overall user experience. This chapter provides a detailed overview of the required
software components integral to the success of the project.

First and foremost, the system leverages the MongoDB database, a NoSQL database that fits
seamlessly into the MERN Stack architecture. MongoDB serves as the primary data storage
solution, handling the vast amounts of user and product data required for recommendation
generation. Its flexible document-based data model allows for efficient storage and retrieval of
diverse data types, including user profiles, purchase histories, and product details.

Node.js, a server-side JavaScript runtime, forms another critical component of the MERN Stack.
It facilitates the development of a robust backend infrastructure, handling server-side logic, and
serving as a bridge between the frontend and the database. The asynchronous, event-driven nature
of Node.js ensures scalability, making it well-suited for handling concurrent user requests and
interactions.
18
Express.js, a minimal and flexible Node.js web application framework, complements Node.js by
simplifying the creation of server-side applications. Express.js streamlines the development of
RESTful APIs, which are crucial for communication between the frontend and backend
components of the recommendation system. This lightweight framework enhances the efficiency
of handling HTTP requests, enabling seamless data transfer and processing.

On the frontend, the system relies on React.js, a JavaScript library for building user interfaces.
React.js facilitates the creation of dynamic and responsive user interfaces, essential for presenting
personalized product recommendations to users. Its component-based architecture allows for the
modular development of the frontend, promoting code reusability and maintainability.

Machine Learning algorithms form the core of the recommendation system, and their integration
is facilitated by Python, a versatile programming language for data science and machine learning.
Python provides a rich ecosystem of libraries, such as scikit-learn and TensorFlow, which are
instrumental in building and training recommendation models. These models analyze user
behavior, purchase history, and product attributes to generate personalized recommendations.

The integration of a recommendation engine, powered by ML algorithms, is a pivotal component

of the system. Collaborative filtering, content-based filtering, and hybrid recommendation
techniques are employed to analyze user preferences and suggest products that align with
individual tastes. The recommendation engine continuously refines its models through iterative
learning, adapting to changing user preferences over time.

Furthermore, the incorporation of a messaging queue system, such as RabbitMQ or Kafka,

enhances the scalability and responsiveness of the recommendation system. This ensures that
recommendation requests and updates are efficiently processed, minimizing latency and delivering
real-time suggestions to users.

In conclusion, the Personalized Recommendation System for an E-Commerce Website using the
MERN Stack and ML is underpinned by a well-defined set of software components. From the
database and server infrastructure to the frontend user interface and machine learning algorithms,

19
each component plays a crucial role in creating a seamless and personalized shopping experience
for users. The synergy of these components forms the foundation of a robust and effective
recommendation system, enhancing user engagement and satisfaction in the dynamic realm of e-
commerce.

4.2 IMPLEMENTATION

The implementation phase of the Personalized Recommendation System for an E-Commerce

Website using the MERN (MongoDB, Express.js, React, Node.js) Stack and Machine Learning
involves a meticulous process that encompasses both front-end and back-end development. This
phase aims to bring the theoretical concepts and design specifications into a fully functional and
user-friendly system. The following paragraphs delve into the key aspects of the implementation
process.

At the heart of the implementation is the development of the back-end infrastructure using Node.js
and Express.js. This involves creating RESTful APIs that handle various functionalities, including
user authentication, product catalog management, and recommendation generation. The
MongoDB database, chosen for its scalability and flexibility, stores user profiles, purchase history,
and product details. The integration of these technologies provides a robust foundation for the
recommendation system to thrive.

Machine Learning (ML) plays a pivotal role in delivering personalized recommendations to users.
Collaborative filtering algorithms, such as matrix factorization or user-item similarity models, are
implemented to analyze user behavior and preferences. The ML models are trained on historical
user interactions and purchase data, allowing the system to learn and adapt to evolving user
preferences over time. The integration of ML into the recommendation system enhances its ability
to provide accurate and relevant product suggestions tailored to individual users.

On the front-end, the React framework is employed to create a dynamic and responsive user
interface. The user experience is paramount in an E-Commerce setting, and React's component-
based architecture facilitates the development of interactive and seamless web pages. The
recommendation engine's outputs are seamlessly integrated into the user interface, ensuring that
recommended products are prominently displayed and easily accessible. This user-centric design
20
enhances engagement and encourages users to explore personalized recommendations.

To enhance the system's performance and responsiveness, asynchronous operations are

implemented using tools like Axios for making asynchronous HTTP requests. This ensures that
the user experience remains smooth even when fetching data from the server or updating the user
interface dynamically. Additionally, state management libraries like Redux may be integrated to
efficiently manage the application's state, ensuring consistency across components and improving
overall responsiveness.

Security is a critical consideration in the implementation phase. User authentication and

authorization mechanisms are implemented to secure user data and ensure that personalized
recommendations are only accessible to the respective users. Encryption protocols, such as
HTTPS, are implemented to protect data transmission between the client and server, safeguarding
sensitive information from potential threats.

Scalability is another key aspect of the implementation, especially in the context of an E-

Commerce website that may experience varying levels of traffic. The architecture is designed to
handle increased loads efficiently, with server clusters and load balancing mechanisms
implemented as necessary. This ensures that the recommendation system remains responsive and
reliable, even during peak usage periods.

Testing is conducted rigorously throughout the implementation phase. Unit testing, integration
testing, and end-to-end testing are employed to identify and rectify bugs or inconsistencies in the
system. User acceptance testing is also crucial to ensuring that the personalized recommendation
system aligns with user expectations and provides a positive user experience.

In conclusion, the implementation of a Personalized Recommendation System for an E-Commerce

Website using the MERN Stack and Machine Learning involves the seamless integration of back-
end technologies, machine learning algorithms, and a user-friendly front-end. The collaborative
efforts of these components result in a robust, scalable, and secure system that not only meets user
expectations but also anticipates and adapts to their evolving preferences.

21
4.3 THEORY HARDWARE AND SOFTWARE REQUIREMENTS

Hardware Requirements:
In the development and deployment of a Personalized Recommendation System for an E-
Commerce Website using the MERN Stack and Machine Learning (ML), the hardware
infrastructure plays a crucial role. The following hardware requirements are essential for the
successful implementation of the project:

Server Infrastructure: A robust server infrastructure is necessary to handle the computational

load associated with ML algorithms and database operations. This includes high-performance
processors, ample RAM, and scalable storage solutions to accommodate the growing dataset.

Storage Systems: Given the data-intensive nature of recommendation systems, storage systems
with high throughput and reliability are indispensable. The use of SSDs or other high-speed storage
technologies is recommended to ensure quick retrieval of user and product data.

Networking Equipment: A high-speed and reliable network infrastructure is essential for

seamless communication between the server, database, and other components. Adequate
bandwidth is crucial for handling concurrent user requests efficiently.

Software Requirements:
The software stack chosen for the development of the Personalized Recommendation System is
pivotal in ensuring compatibility, scalability, and efficiency. The following software components
are integral to the project:

Operating System: A Linux-based operating system, such as Ubuntu or CentOS, is preferred for
its stability, security, and compatibility with the MERN (MongoDB, Express.js, React, Node.js)
stack.

Web Server: Node.js will serve as the web server, providing a non-blocking, event-driven
architecture suitable for handling numerous concurrent connections efficiently.
22
Database Management System (DBMS): MongoDB, a NoSQL database, is well-suited for
handling large volumes of unstructured data, making it an ideal choice for an E-Commerce
Recommendation System.

Programming Language: The MERN stack relies on a combination of programming languages

for both server-side and client-side development:

JavaScript: As a versatile language, JavaScript is used for both frontend (React) and backend
(Node.js) development, ensuring a consistent and seamless user experience.

Data Visualization Tools: To gain insights from the recommendation system and present them in
a comprehensible manner, data visualization tools like D3.js or Chart.js can be integrated. These
tools enhance the interpretability of complex data patterns and user preferences.

Integrated Development Environment (IDE): An integrated development environment

streamlines the coding and debugging process. Visual Studio Code, a lightweight and feature-rich
IDE, is an excellent choice for MERN stack development, offering code highlighting, IntelliSense,
and Git integration.

Version Control: Effective version control is crucial for collaborative development and code
management. Git, with platforms like GitHub or GitLab, will be employed to track changes,
manage branches, and facilitate collaboration among team members.

Security Software: Ensuring the security of user data and the recommendation algorithm is
paramount. Security software, including encryption protocols, firewalls, and intrusion detection
systems, will be implemented to safeguard against potential threats.

Model Deployment Tools: Efficient deployment tools are essential for releasing updates and new
features. Docker, coupled with orchestration tools like Kubernetes, enables containerization and
automates the deployment process, ensuring consistency across various environments.

23
Theory Requirements: Firstly, a profound grasp of Machine Learning algorithms is paramount.
The foundational concepts of supervised and unsupervised learning, as well as their application in
recommendation systems, are crucial. Supervised learning models, such as regression and
classification algorithms, are vital for predicting user preferences based on historical data.
Unsupervised learning, including clustering techniques like k-means, helps in identifying patterns
and grouping similar items or users.
Moreover, knowledge of collaborative filtering is essential for building a recommendation system
that leverages user behaviour and preferences. Collaborative filtering methods, whether user-based
or item-based, rely on the principle of user-item interaction patterns to generate accurate
suggestions. Understanding the strengths and limitations of collaborative filtering algorithms is
pivotal for choosing the most suitable approach for the E-Commerce platform.
In addition to Machine Learning, a solid foundation in the principles of the MERN Stack is
indispensable. Proficiency in MongoDB, a NoSQL database, is essential for efficiently managing
and storing vast amounts of user and product data. Express.js, as the backend framework,
facilitates the creation of robust and scalable APIs, ensuring seamless communication between the
frontend and backend components. React, the frontend library, plays a pivotal role in providing an
interactive and dynamic user interface, enhancing the overall user experience. Node.js serves as
the runtime environment for executing server-side JavaScript code, ensuring smooth server
operations.
Understanding the importance of data preprocessing is another critical theoretical aspect. Clean
and well-organized data is imperative for training accurate Machine Learning models. Techniques
such as data cleaning, normalization, and feature engineering play a vital role in preparing the
dataset for optimal model performance.
In conclusion, the successful development of a Personalized Recommendation System for an E-
Commerce Website using the MERN Stack and ML hinges on a deep understanding of Machine
Learning algorithms, collaborative filtering, the MERN Stack components, data preprocessing,
and the challenges inherent in recommendation systems. Armed with this theoretical knowledge,
developers can design and implement a recommendation system that enhances user engagement
and satisfaction on the E-Commerce platform.

24
 CHAPTER 5

RISK ASSESSMENT

Risk assessment for an e-commerce website with an integrated recommendation system is

imperative to safeguard both businesses and customers from potential threats. The integration of
a recommendation system introduces complexities, as inaccurate or biased recommendations may
misguide customers. Additionally, the collection and utilization of user data for personalization
raise concerns about privacy and regulatory compliance. Security risks, such as data breaches and
unauthorized access, pose a significant threat, especially considering the sensitive customer
information involved in online transactions. Scalability challenges and performance optimization
issues can impact user experience and hinder the website's ability to handle increasing traffic.
Ensuring seamless integration with third-party services, such as payment gateways and inventory
management, is crucial to prevent disruptions. Furthermore, the potential for legal and compliance
issues, including adherence to data protection laws, necessitates careful risk mitigation strategies.
Proactive risk assessment and mitigation efforts are essential to create a robust e-commerce
platform that not only enhances user experience but also safeguards against potential pitfalls.

5.1 SECURTY RISK

RISK:

E-commerce websites deal with sensitive customer information, including personal details and
payment data. Ensuring robust security measures (like encryption, secure APIs, and proper
authentication) is crucial to prevent data breaches and unauthorized access.

IMPACT:

Unauthorized access, data breaches, and compromised customer information.

MITIGATION STRATEGY:

Implement secure coding practices, use encryption for sensitive data, conduct regular security
audits, and stay informed about the latest security threats.
25
5.2 SCALABILTY RISK

RISK:

As the user base grows, the website may experience scalability challenges. Adequate planning and
design considerations for scalability are essential to handle increasing traffic and transactions
without compromising performance.

IMPACT:

As the user base grows, the website may experience scalability challenges. Adequate planning and
design considerations for scalability are essential to handle increasing traffic and transactions
without compromising performance.

MITIGATION STRATEGY:

Design the system for scalability from the start, use scalable infrastructure (e.g., cloud services),
and regularly assess and optimize performance.

5.3 PERFORMANCE OPTIMIZATION RISK

RISK:

Optimizing the performance of an e-commerce website is crucial for providing a smooth user
experience. Large product catalogs, heavy traffic, and complex features can impact the website's
speed. Caching mechanisms, code splitting, and proper database indexing are some strategies to
address performance issues.

IMPACT:

Slow website speed, leading to a decrease in user engagement.

MITIGATION STRATEGY:

Utilize caching mechanisms, optimize code and assets, employ content delivery networks (CDNs),
and monitor performance regularly.

26
5.4 INTEGERATION COMPLEXITY RISK

RISK:

Integrating various components and third-party services, such as payment gateways, shipping
APIs, and inventory management systems, can be complex. Compatibility issues and API changes
may pose challenges during integration.

IMPACT:

Delays in project timelines, potential disruptions in services, and increased development costs.

MITIGATION STRATEGY:

Thoroughly plan integrations, use well-documented APIs, and keep abreast of changes in third-
party services.

5.5 USER EXPERIENCE DESIGN RISK

RISK:

Designing an intuitive and user-friendly interface is critical for e-commerce success. Failing to
prioritize user experience may lead to decreased customer satisfaction and conversion rates.

IMPACT:

Low conversion rates, dissatisfied customers, and reduced competitiveness.

MITIGATION STRATEGY:

Prioritize user-centric design, conduct usability testing, and continuously gather user feedback for
iterative improvements.

5.6 RECOMMENDATION SYSTEM CHALLENGES

RISK:

27
Implementing an effective recommendation system involves dealing with large datasets,
algorithms, and continuous improvement. The challenge lies in creating accurate and relevant
recommendations, as well as managing and updating recommendation models.

IMPACT:

Inaccurate or irrelevant recommendations, leading to decreased user engagement.

MITIGATION STRATEGY:

Continuously refine and update recommendation algorithms, employ A/B testing for new
algorithms, and allow users to provide feedback on recommendations.

5.7 PERSONALIZED AND PRIVACY CHALLENGES

RISK:

While recommendation systems aim to provide personalized experiences, there is a fine line
between helpful suggestions and concerns about user privacy. Balancing personalization with
privacy regulations and user consent is crucial.

IMPACT:

Loss of user trust, potential legal issues, and increased scrutiny.

MITIGATION STRATEGY:

Clearly communicate privacy practices, provide transparent opt-in/opt-out options for

personalization, and comply with relevant privacy regulations.

5.8 PERSONALIZED AND PRIVACY CHALLENGES

RISK:

28
While recommendation systems aim to provide personalized experiences, there is a fine line
between helpful suggestions and concerns about user privacy. Balancing personalization with
privacy regulations and user consent is crucial.

IMPACT:

Loss of user trust, potential legal issues, and increased scrutiny.

MITIGATION STRATEGY:

Clearly communicate privacy practices, provide transparent opt-in/opt-out options for

personalization, and comply with relevant privacy regulations.

29
 CHAPTER 6

APPLICATION ARCHITECTURE
In the realm of contemporary e-commerce, the integration of data-driven product
recommendations stands out as a pivotal advancement in enhancing user experience and driving
sales. This project module revolves around the development and implementation of a Next-Gen
E-commerce system, leveraging cutting-edge technologies and innovative design strategies to
deliver personalized and relevant product suggestions to users. The core focus lies in harnessing
the power of big data analytics, machine learning algorithms, and artificial intelligence to analyze
user behavior, preferences, and purchase history. The project aims to design an intelligent
recommendation engine that adapts and evolves with user interactions, creating a seamless and
personalized shopping journey. By delving into the intricacies of user data, this module seeks to
revolutionize the conventional e-commerce landscape by providing a more intuitive and satisfying
shopping experience, ultimately fostering increased customer engagement and conversion rates.
The design considerations encompass not only the robust functionality of the recommendation
engine but also the user interface and experience, ensuring a visually appealing and user-friendly
platform that aligns with the evolving expectations of modern online consumers.

Figure 6.1: Use case- Diagram

30
Furthermore, the project module focuses on the intricate details of the Next-Gen E-commerce
system, incorporating meticulous designs that emphasize the importance of user interaction and
engagement. The recommendation engine is engineered to adapt in real-time, constantly learning
from user feedback and refining its suggestions to cater to evolving tastes and preferences. This
dynamic approach ensures that customers are presented with a curated selection of products that
align not only with their past behavior but also with current trends and emerging interests. In
addition to the backend algorithms, the frontend design of the e-commerce platform is meticulously
crafted to enhance the overall user experience. Intuitive navigation, visually appealing product
displays, and seamless integration of the recommendation engine into the user interface are key
design considerations. A responsive and mobile-friendly design is implemented to cater to the
growing trend of mobile commerce, ensuring that users have a consistent and enjoyable experience
across various devices. The project also explores the incorporation of emerging technologies such
as augmented reality for virtual try-ons or product visualization, adding an extra layer of innovation
to the shopping experience. By combining sophisticated data analysis with thoughtful design, this
project aims to set a new standard for e-commerce platforms, creating a synergistic relationship
between cutting-edge technology and user-centric design.

6.1 Phases of project

There are many features of this project. So the project is currently being developed in eight phases.

Phase 1: Project Initiation

 Define the project objectives, goals, and scope.
 Identify stakeholders and establish communication channels.
 Conduct a feasibility study to assess the viability of the project.
Phase 2: Requirement Gathering
 Work with stakeholders to gather detailed functional and non-functional requirements.
 Define the features and functionalities of the e-commerce website and recommendation
system.
Phase 3: System Design
 Create a high-level system architecture, outlining the components and their interactions.
31
  Design the database schema for storing product information, user data, and recommendation
models.
 Define the user interface and user experience (UI/UX) design.

Phase 4: Technology Stack Selection

 Choose appropriate technologies for the development of the e-commerce platform and
recommendation system.
 Consider factors such as scalability, security, and integration capabilities.

Phase 5: Development
 Implement the e-commerce website, including user registration, product catalog, shopping
cart, and checkout functionalities.
 Develop the recommendation system, integrating machine learning algorithms for
personalized suggestions.
 Implement secure payment gateways and other essential features.

Phase 6: Testing
 Conduct unit testing, integration testing, and system testing to ensure the functionality and
reliability of the entire system.
 Perform security testing to identify and address potential vulnerabilities.

Phase 7: Deployment
 Deploy the e-commerce website and recommendation system to a staging environment for
final testing.
 After successful testing, deploy the system to the production environment.

Phase 8: Optimization and Continuous Improvement

 Analyze user behavior and system performance to optimize the recommendation algorithms.
 Gather user feedback and continuously improve the e-commerce website based on evolving
market trends.

32
 CHAPTER 7

SOFTWARE FEATURES

1. User Authentication and Profiles:

 User authentication is the foundation, allowing users to create accounts and log
in securely.
 Detailed user profiles enhance personalization by storing preferences and order
history.
 Secure login methods and password protection ensure data confidentiality.
2. Intuitive Product Catalog:
 A well-organized and easily searchable product catalog simplifies the user's
navigation.
 Implementing product categories and filters provides a seamless shopping
experience.
 A visually appealing layout enhances user engagement and encourages
exploration.
3. Efficient Shopping Cart and Checkout:
 Streamlined cart functionality with easy addition and removal of items.
 A secure and straightforward checkout process with multiple payment options
boosts user satisfaction.
 Order confirmation and tracking options keep users informed about their
purchases.
4. Smart Search Functionality:
 A robust search function, coupled with filters, ensures users quickly find what
they're looking for.
 Auto-suggestions and spell-check enhance the search experience, reducing
friction for users.
5. Dynamic Recommendation System:
 Personalized product recommendations based on user behavior and preferences
elevate the user experience.
33
  Employing collaborative filtering or content-based recommendation algorithms
refines suggestions.
6. Engaging Product Reviews and Ratings:
 User-generated product reviews and ratings build trust and aid in informed
decision-making.
 Integration with social media for sharing reviews enhances product visibility and
credibility.
7. Wishlist and Notification Alerts:
 Wishlist empower users to curate desired items for future purchase.
 Notification alerts for price drops or restocked items keep users engaged and
informed.
8. Responsive Design for Accessibility:
 A mobile-friendly design ensures a seamless shopping experience across various
devices.
 Responsive design contributes to increased accessibility, accommodating diverse
user preferences.
9. Robust Security Measures:
 SSL encryption secures transactions and protects sensitive user data.
 Multi-factor authentication adds an extra layer of security, enhancing user
account protection.

34
 CHAPTER 8

PROJECT SNAPSHOTS

Fig 8.1 Login page

35
Fig 8.2 Home page

36
 REFERENCE

[1] Caiwen Li, Iskandar Ishak, Hamidah Ibrahim, Maslina Zolkepli, Fatimah Sidi, Caili Li, "Deep
Learning-Based Recommendation System: Systematic Review and Classification," IEEE Journals link:
https://ieeexplore.ieee.org/document/10274963/

[2] Amal Albalawi & Saad T Alharbi, " E-Commerce Recommender Systems: A Comprehensive
Survey," source: https://www.researchgate.net/publication/373458583_E-
Commerce_Recommender_Systems_A_Comprehensive_Survey/

[3] Sabina-Cristina Necula & Vasile-Daniel Pavaloaia, " AI-Driven Recommendations: A Systematic
Review of the State of the Art in E-Commerce," MDPI Journals link: https://www.mdpi.com/2076-
3417/13/9/5531/

[4] Chia-Huei Wu , “e-Commerce Personalized Recommendation Based on Machine Learning

Technology,” Hindwai Journals link: https://www.hindawi.com/journals/misy/2022/1761579/

[5] Brown, A. C., Harris, M. P., Turner, D. W. (2020). "Advances in Secure Payment Systems for
E-commerce," IEEE Transactions on Information Systems. link: https://www.mdpi.com/2073-
431X/9/3/66

[6] Taylor, R. M., Johnson, J. A., Lewis, K. L. (2020). "Social Media Integration in E-commerce,"
Proceedings of Social Commerce Conference link:
https://www.researchgate.net/publication/338855010_A_Study_on_Connecting_Social_media_to_
E-Commerce_Mechanism

[7] Chen, E. S., Lee, D. M. (2022). "Enhancing E-commerce Recommendations Through Deep
Learning: A Case Study on Fashion Retail," IEEE Transactions on Neural Networks and Learning
Systems. link: https://www.mdpi.com/2073-431X/9/3/66

[8] Patel, J. R., Chen, L. M. (2022). "Addressing Cold Start in Recommendation Systems: A
Content-Based Approach," Proceedings of the International Conference on Data Mining. link:
https://www.mdpi.com/1999-4893/16/4/182
37