SRM INSTITUTE OF SCIENCE AND TECHNOLOGY

Ramapuram, Chennai – 600 089

SCHOOL OF COMPUTER SCIENCE AND ENGINEERING
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

18CSP109L-MAJOR PROJECT
First Review

BATCH NUMBER : 17
Product Review Evaluation Using LLMs

Team Members Supervisor

Vignesh R (RA2111027020066) NAME: Mr.A.Thiruneelakandan,
Ashwath S (RA2111027020074) Assistant Professor,
Rahul Bundele (RA2111027020078) Department of CSE

Date: 22-02-2025 Slide

Number: 18
 Agenda
• Abstract
• Literature survey
• Scope and Motivation
• Objectives
• Novel idea
• Block diagram of proposed system
• Tools to be used
• Applications
• References

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Abstract
This project, titled "Product Review Evaluation Using LLMs," focuses on
building an interactive web application where users can input a product
name to retrieve and analyze customer reviews from Amazon. The system
employs web scraping techniques to gather reviews and utilizes the
ChatGPT API for sentiment analysis. The output comprises a
comprehensive list of product pros and cons, along with the count of
mentions for each aspect. This innovative approach aims to assist
customers in making informed purchasing decisions by providing
summarized insights from extensive reviews.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Scope And Motivation
● Scope: The project addresses the challenge of sifting through
hundreds or thousands of online reviews to find meaningful insights.
By automating sentiment analysis and providing a structured
summary, the project benefits both customers and businesses. The
platform is scalable, with the potential to include multiple e-
commerce websites and languages in the future.

● Motivation: In the era of e-commerce, reviews significantly influence

buying decisions. However, customers often face information
overload. By leveraging AI and web technologies, this project
reduces the cognitive load and time spent on research, ensuring a
better shopping experience.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Introduction
• In today’s digital marketplace, customer reviews play a crucial role in
shaping purchasing decisions. With millions of reviews available for various
products, extracting meaningful insights from them can be a daunting task.
Traditional review analysis methods often require extensive manual effort or
rely on simplistic sentiment classification, which may fail to capture
nuanced opinions.
• This project, Product Review Evaluation Using LLMs, aims to bridge this
gap by leveraging advanced language models and automation to streamline
review analysis. By integrating web scraping techniques and the ChatGPT
API for sentiment analysis, the system efficiently gathers and processes
customer feedback. The resulting insights include a detailed breakdown of
product pros and cons, along with the frequency of mentions for each aspect.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Literature Survey
Journal/Conference
S.No. Title of the Paper Year Inferences
Name
The study examines how LLMs like
GPT-3.5 and LLaMA-2, along with
LLMs in e- NLP models BERT and RoBERTa,
commerce: A Natural Language analyze customer satisfaction in e-
comparative analysis 2024 Processing Journal 6 commerce through product reviews. It
01 of GPT and LLaMA evaluates their performance before and
models in product after fine-tuning, aiming to enhance
review evaluation sentiment analysis and bridge the gap
between emotion detection and
contextual understanding, contributing
to better e-commerce strategies.

Extracting marketing
information from product User-generated content (UGC)
reviews: a comparative contains customer opinions which
study of latent semantic can be used to hear the voice of
02 2023 Computer , Materials &
analysis and probabilistic Continua customers. This information can be
latent semantic analysis useful in market surveillance, digital
innovation, or brand improvisation

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Literature Survey
Journal/Conference
S.No. Title of the Paper Year Inferences
Name
Sentiment analysis via artificial
intelligence, i.e., machine learning
and large language models (LLMs),
Sentiment Analysis of Big Data Cognitive is a pivotal tool that classifies
Product Reviews Using 2024 Computing sentiments within texts as positive,
03 Machine Learning and negative, or neutral. It enables
Pre-Trained LLM computers to automatically detect and
interpret emotions from textual data,
covering a spectrum of feelings
without direct human intervention

We performed a sentiment analysis on

Amazon’s fashion review datasets
using the VADER, BERT, and
GooglePaLM models, respectively,
Customer Sentiments in 2024 Analytics and compared the results based on
04 Product Reviews: A evaluation metrics such as precision,
Comparative Study with recall, accuracy correct positive
GooglePaLM prediction, and correct negative
prediction.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Objectives
• To develop a user-friendly web application to input product names.

• To implement web scraping to collect product reviews from Amazon.

• To use ChatGPT API for sentiment analysis to identify pros and cons.

• To present results in an intuitive format, with counts for pros and cons.

• To ensure the system is scalable and can handle large volumes of data
efficiently.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Problem Statement
With the growing number of online shopping platforms, customers often
struggle to make informed purchasing decisions due to the
overwhelming volume of product reviews. Manually reading and
analyzing reviews is time-consuming and inefficient. There is a need
for an automated system that can extract, analyze, and summarize
product reviews to provide users with clear insights into the pros and
cons of a product.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

Block Diagram Of The Proposed System

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Proposed-Novel Idea
• The uniqueness of this project lies in combining web scraping with
advanced LLMs for real-time sentiment analysis of product reviews.

• Unlike existing platforms that only aggregate reviews, this system

summarizes reviews into actionable insights, presenting pros and cons
with frequency counts.

• This approach saves time for customers and offers businesses valuable
insights into product performance.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Proposed Module
• Product Input Module: Takes the product name as input from the
user.

• Server Module: Stores and manages input product details.

• Web Scraping Module: Scrapes product-related data from online

sources.

• Fetch Products Module: Retrieves all relevant products

matching the search.

• Fetch Reviews Module: Collects all available reviews for the

selected product.
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS
 Proposed Module
• Web Scraping Module: Scrapes product-related data from online
sources.

• Fetch Products Module: Retrieves all relevant products

matching the search.

• Fetch Reviews Module: Collects all available reviews for the

selected product.
 Proposed Module
Natural Language Processing (NLP) Module:

• GPT Model Integration: Uses GPT-based models for

sentiment classification.

• Hugging Face Model Integration: Uses pre-trained models

from Hugging Face.

• NLTK Processing: Uses NLTK for tokenization, sentiment

scoring, and preprocessing.
 Module 1- Outcome
• Captures the product name entered by the user and sends it to the
server for processing.

• tores the product details and manages requests for retrieving relevant
data.

• Extracts product information and reviews from e-commerce websites

or relevant online sources.

• Retrieves a list of products matching the entered product name.

• Collects all user reviews related to the selected product from different
sources.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 Software & Hardware Requirements
•Software Requirements:
Operating System: Windows 10/11, macOS, or Linux (Ubuntu
recommended).
Programming Language: Python 3.7 or later.
Libraries and Frameworks: BeautifulSoup, Selenium, OpenAI API
(ChatGPT), NumPy, Pandas, Matplotlib, Seaborn.
Web Framework: Django.
Development Environment: Jupyter Notebook, PyCharm, or VS Code.
•Hardware Requirements:
Processor: Intel i5/i7 or AMD Ryzen 5/7 (or equivalent).
RAM: 16 GB or more.
GPU: NVIDIA GPU with CUDA support (e.g., GTX 1060 or higher) for
faster model training.
Storage: SSD with at least 500 GB capacity for storing datasets and
models.
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS
 References
• Konstantinos I. Roumeliotis, Nikolaos D. Tselikas, Dimitrios K.
Nasiopoulos. 2024. LLMs in e-commerce: A comparative analysis of GPT
and LLaMA models in product review evaluation.

• Ahmad, S.N., Laroche, M., 2023. Extracting marketing information from

product reviews: A comparative study of latent semantic analysis and
probabilistic latent semantic analysis. J. Market.

• Pawanjit Singh Ghatora 1 , Seyed Ebrahim Hosseini 1,* , Shahbaz Pervez

1 , Muhammad Javed Iqbal 2 and Nabil Shaukat, 2024. Sentiment
Analysis of Product Reviews Using Machine Learning and Pre-Trained
LLM.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS

 References
• Olamilekan Shobayo 1 , Swethika Sasikumar , Sandhya Makkar and
Obinna Okoyeigbo 2024.Customer Sentiments in Product Reviews: A
Comparative Study with GooglePaLM.

• Brown, T.B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P.,
Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-
Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D.M.,
Wu, J., Winter, C., 2020. Language models are few-shot learners. Adv.
Neural Inf. Process. Syst. 2020-December.

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING - BIG DATA ANALYTICS