Sentiment Analysis on Amazon Reviews

A Flask-based API that analyzes sentiment in Amazon product reviews using machine learning. It processes text data, applies TF-IDF vectorization, and uses a Logistic Regression model to classify reviews as positive, negative, or neutral.

Features

• Preprocesses and cleans text data
• Uses TF-IDF vectorization for feature extraction
• Logistic Regression for sentiment classification
• Simple API endpoint for predictions

How to Run

Step 1: Clone the Repository

git clone https://github.com/nipungoel24/Amazon-Review-Sentiment-Analysis.git
cd Amazon-Review-Sentiment-Analysis

Step 2: Create and Activate a Conda Environment

conda create -n amazonreview python=3.10
conda activate amazonreview

Step 3: Install Dependencies

pip install -r requirements.txt

Step 4: Run the Flask API

flask --app api.py run

Step 5: Access the API

Once the app starts, it will run on port 5000 by default:

http://localhost:5000

API Endpoints

Method	Endpoint	Description
POST	/predict	Accepts a review as input and returns its sentiment (positive/negative).

Future Enhancements

• Implement deep learning models for better accuracy
• Deploy the API on cloud platforms
• Improve text preprocessing techniques

Contributions

Feel free to fork the repo and open a pull request with your improvements!