Recommendation Model

An AI/ML recommendation system built w/ Python, that has collaborative filtering, matrix factorization, and content-based filtering.

Tech Stack

• Python 3.9+
• FastAPI (REST API framework)
• scikit-learn (ML algorithms)
• Pandas & NumPy (data processing)

Quick start

# Install dependencies
pip install -r requirements.txt

# Generate sample data
python3 generate_data.py

# Test models
python3 models.py

# Start API server
python3 api.py

Visit http://localhost:8000/docs for interactive API documentation.

Features

• Collaborative Filtering (User & Item-based)
• Matrix Factorization (SVD)
• Content-Based Filtering
• RESTful API with Swagger docs
• Real-time recommendations

API Endpoints

• POST /recommend - get personalized recommendations
• GET /movies - browse movies
• GET /health - health check

Algorithms

1. Collaborative Filtering: Finds similar users/items using cosine similarity
2. Matrix Factorization: SVD-based latent factor model
3. Content-Based: Recommends based on genre and features

Project Structure

movie-recommendation-system/
├── api.py                  # FastAPI server w/ recommendation endpoints
├── models.py              # ML models (Collaborative Filtering, Matrix Factorization)
├── generate_data.py       # synthetic dataset generator
├── test_system.py         # testing script for validation
├── requirements.txt       # python dependencies
├── README.md             # project documentation
├── .gitignore            # Git ignore rules
│
├── movies.csv            # generated movie data (100 movies)
├── users.csv             # generated user data (50 users)
└── ratings.csv           # generated ratings data (~1000 ratings)

File Descriptions

• api.py: FastAPI REST API server with endpoints for recommendations, movies, and health checks
• models.py: Implementation of recommendation algorithms
  • SimpleRecommender: Item-based collaborative filtering
  • MatrixFactorization: SVD-based factorization model
• generate_data.py: Creates synthetic movie dataset with realistic rating patterns
• test_system.py: Automated testing for data, models, and API
• requirements.txt: All Python package dependencies

Screenshots