Sentiment Analysis of Online Product Reviews for E-Commerce System

ABSTRACT:

In the modern era of e-commerce, online product reviews play a crucial role in influencing customer decisions. Buyers often rely on user generated reviews to assess product quality, reliability, and overall satisfaction. However, due to the vast volume of reviews available, manually analyzing them becomes impractical. Sentiment Analysis of Online Product Reviews for E-Commerce System is developed to address this challenge by leveraging Java for backend development, JSP, CSS, and JavaScript for the front-end, and MySQL for database management. This system classifies product reviews into positive and negative sentiments using a predefined dataset of sentiment words, helping users make informed purchasing decisions.

The need for such a system arises due to the increasing number of online transactions and the potential influence of fake, misleading, or biased reviews. Traditional e-commerce platforms provide a platform for customer feedback but lack an intelligent mechanism to classify and filter reviews based on sentiment. By implementing sentiment analysis, the system ensures that users can easily distinguish between positive and negative opinions, thus improving the credibility and transparency of online shopping experiences. Furthermore, businesses can utilize sentiment classification to gain insights into customer preferences and enhance their services.

By implementing sentiment analysis, this system enhances the trustworthiness of online reviews, helping users make better purchasing decisions while enabling businesses to improve their products based on customer feedback. The automated classification of reviews also helps in filtering out misleading or biased opinions, ensuring a more reliable e-commerce experience.

PROJECT OUTPUT VIDEO:

EXISTING SYSTEM:

The existing system for managing online product reviews in e-commerce platforms primarily relied on manual review submission and display mechanisms without any automated classification. Customers could post their feedback on products, and these reviews were displayed in a chronological order. Users had to read through multiple reviews to get an overall idea of the product quality and customer satisfaction.
In the existing system approach, the system allowed users to register, log in, and browse through different products based on categories. Each product page displayed details such as product name, price, description, and images, along with customer reviews submitted by other buyers. These reviews were stored in the system’s database and retrieved for display when a user accessed the respective product page.
The system also maintained search functionality, enabling users to find products by entering specific keywords. Additionally, some platforms included basic filtering options, allowing users to sort reviews based on date, rating, or helpfulness as marked by other users. The database stored user-generated content without any sentiment-based classification, and all reviews were displayed as they were posted by customers.
For administrative tasks, the system provided an interface where the admin could manage product categories, add new products, and monitor user activities. The admin had access to all reviews and could manually verify or moderate them if needed. The system also maintained logs of users’ search history and product interactions, helping admins track user engagement.
Thus existing system effectively provided a platform for user reviews and product feedback, allowing customers to share their experiences. However, it lacked an automated approach to classify and analyze sentiment, requiring users to manually go through multiple reviews to form an opinion about a product.

DISADVANTAGES OF EXISTING SYSTEM:

Lack of Sentiment Classification: The earlier system displayed all customer reviews in a simple list format without classifying them as positive or negative. Users had to manually read multiple reviews to determine the overall sentiment about a product, making the decision-making process time-consuming.
Difficulty in Analyzing Reviews: Since reviews were stored as raw text without sentiment-based categorization, users found it challenging to quickly assess product quality. There was no automated way to summarize customer feedback, leading to a lack of clarity in understanding overall customer satisfaction.
No Review-Based Insights: The system did not generate statistical or graphical representations of product reviews. Users and administrators could not view insights like review trends, customer satisfaction rates, or product popularity based on sentiment analysis.
Time-Consuming Review Moderation: Since there was no automated filtering of reviews based on sentiment, admins had to manually verify and monitor user feedback. This increased administrative workload and made it difficult to manage large volumes of reviews efficiently.
No Automated Review Validation: The system allowed all submitted reviews to be stored without checking whether they contained meaningful content. This led to situations where irrelevant, fake, or misleading reviews could be posted, affecting the reliability of customer feedback.
No Domain-Based Analysis of Reviews: In the existing system, there was no feature to analyze reviews based on product domains (e.g., electronics, fashion, home appliances). As a result, administrators and business owners lacked insights into which product categories were receiving more positive or negative feedback.
These limitations in the existing system highlight the need for a sentiment-based classification approach, enabling users to quickly identify positive and negative reviews, making the review system more effective and user-friendly.

PROPOSED SYSTEM:

The proposed system, “Sentiment Analysis of Online Product Reviews for E-Commerce System” introduces an advanced review classification mechanism that automatically categorizes user reviews into positive and negative sentiments. This system is designed to improve the efficiency of handling product reviews by integrating sentiment analysis techniques with an e-commerce platform.
The system is developed using Java for backend processing, JSP for the frontend interface, and MySQL as the database to store user reviews and product information. It uses a predefined dataset of positive and negative words to analyze and classify reviews automatically. When a user submits a review for a product, the system processes the text and assigns it a sentiment category based on the occurrence of positive and negative keywords.
The system consists of two primary entities: User and Admin. A new user must first register by providing Name, Email ID, Password, Gender, Date of Birth, Contact Number, and Location. After successful login, the user gains access to three functionalities:
- Search Products – Users can search for a product by entering its name. If the product exists, its details including Product ID, Domain, Product Image, Name, Price, and Description are displayed. Additionally, all reviews posted by other users are visible, and the user can submit their own review in the “Your Review” section.
- View My Search History – Users can view their previously searched products, along with details such as Serial Number, Product Name, Date, and Time.
- View All Products – Users can browse all available products without using the search option.
The Admin has several key functionalities for managing the platform, including:
- Add New Category – The admin can add new product categories (e.g., Mobile, Laptop, etc.) and view the existing categories.
- Add Product – Admin can add new products under specific categories by providing Category Name, Product Name, Price, Description, and Product Image.
- View Products – Displays all added products with Product ID, Category, Name, Price, Description, and Image.
- Users Search History – Admin can track all users’ search histories, including User ID, Username, Product Name, Date, and Time.
- Positive Sentiment Reviews – Reviews containing positive sentiment words (from a predefined dataset) are classified and displayed along with Product ID, Product Name, and Review Content. The system assigns higher reputation values to reviews with more positive words.
- Negative Sentiment Reviews – Similarly, reviews with negative sentiment words are classified and displayed based on a dataset containing negative keywords.
- All Reviews – The admin can view all positive and negative reviews together.
- Graphical Analysis – The system provides a dynamic graph displaying the review count per domain, helping administrators analyze user engagement and product popularity.
The system is designed to handle large volumes of product reviews efficiently by integrating sentiment classification techniques with a structured database. It eliminates the need for manual review classification and provides a centralized platform for users and administrators to manage and analyze product feedback in a structured and meaningful way.

ADVANTAGES OF PROPOSED SYSTEM:

The “Sentiment Analysis of Online Product Reviews for E-Commerce System” offers several advantages over traditional review management approaches by incorporating automated sentiment analysis and structured data processing. The key benefits of the proposed system are as follows:
Automated Sentiment Classification: The system automatically classifies reviews into positive and negative categories based on predefined datasets, eliminating the need for manual review moderation.
This ensures faster and more accurate classification of user feedback.
Improved User Experience: Users can easily search for products and view categorized reviews, helping them make informed purchasing decisions.
The ability to access previously searched products enhances convenience and usability.
Efficient Review Management: The admin can monitor and analyze customer feedback effectively through separate sections for positive and negative reviews.
This helps businesses gain insights into customer satisfaction and product performance.
Graphical Representation of Data: The system generates dynamic graphs to display review counts per domain, providing a visual representation of customer feedback trends.
This enables the admin to make data-driven decisions regarding product improvements.
Time-Saving and Scalable: Unlike manual methods, which require significant time and effort, the proposed system processes reviews instantly and can handle large volumes of data efficiently.
It is designed to be scalable, making it suitable for small and large e-commerce platforms.
Enhanced Decision-Making for Admins: By analyzing user search history and sentiment trends, the admin can identify popular products and address negative feedback promptly.
This helps in improving customer satisfaction and business growth.
Secure and Structured Data Management: The system ensures secure user authentication, preventing unauthorized access to reviews and product data.
MySQL database is used for structured data storage, ensuring efficient retrieval and management of user interactions and product details.
User-Friendly Interface: The system’s interface is built using JSP, CSS, and JavaScript, ensuring a responsive and intuitive design that enhances usability for both users and admins.
Overall, the proposed system enhances customer trust, business intelligence, and e-commerce efficiency by providing a structured, automated, and scalable sentiment analysis solution for online product reviews.

SYSTEM REQUIREMENTS:

HARDWARE REQUIREMENTS:

System : Pentium i3 Processor.
Hard Disk : 20 GB.
Monitor : 15’’ LED.
Input Devices : Keyboard, Mouse.
Ram : 4 GB.

SOFTWARE REQUIREMENTS:

Operating system : Windows 10/11.
Coding Language : Java.
Frontend : JSP, CSS, JavaScript.
JDK Version : JDK 23.0.1.
IDE Tool : Apache Netbeans IDE 24.
Tomcat Server Version : Apache Tomcat 9.0.84
Database : Mysql

Java Projects

Sentiment Analysis of Online Product Reviews for E-Commerce System