Product Analysis and Recommendation System

This project combines BERT-based Named Entity Recognition (NER), sentiment analysis, and product search capabilities to provide intelligent product recommendations based on user descriptions and online sentiment.

Features

• Product description analysis using BERT NER
• Amazon product search integration
• Reddit forum sentiment analysis
• BERT-based sentiment analysis
• Intelligent product recommendations

Setup

1. Install dependencies:

pip install -r requirements.txt

2. Create a .env file with the following credentials:

REDDIT_CLIENT_ID=your_reddit_client_id
REDDIT_CLIENT_SECRET=your_reddit_client_secret
REDDIT_USER_AGENT=your_app_name
AMAZON_ACCESS_KEY=your_amazon_access_key
AMAZON_SECRET_KEY=your_amazon_secret_key
AMAZON_MARKETPLACE_ID=your_marketplace_id

3. Run the application:

python main.py

Project Structure

• main.py: Main application entry point
• src/
  • ner/: BERT NER implementation
  • sentiment/: Sentiment analysis module
  • search/: Amazon product search integration
  • reddit/: Reddit API integration
  • recommendation/: Product recommendation system
• config/: Configuration files
• models/: Saved model files
• data/: Data storage

Usage

1. Enter a product description
2. The system will:
   • Extract key product characteristics using BERT NER
   • Search for matching products on Amazon
   • Analyze Reddit discussions about similar products
   • Perform sentiment analysis
   • Provide product recommendations with sentiment scores

Requirements

• Python 3.8+
• See requirements.txt for package dependencies