🎬 Movie Recommendation System

Smart, Genre-Aware, Rating-Optimized Movie Discovery

This project is a content-based movie recommendation system that intelligently suggests movies based on:

• 🎭 Genre similarity
• ⭐ Movie ratings
• ⚖️ Weighted ranking (similarity + quality)

Unlike naive keyword-based recommenders, this system:

• Ignores misleading title similarity
• Scales safely without building massive similarity matrices

It offers two powerful recommendation modes through a modern web UI.

• Input any movie from the dataset
• Finds genre-similar movies
• Ranks them using a weighted score
• Handles typos with fuzzy matching suggestions

• Select one or multiple genres
• Uses AND-logic (all selected genres must match)
• Returns top-rated movies
• Random sampling ensures non-repetitive results

• Single TF-IDF model trained on genres
• On-demand similarity computation (memory-safe)
• Weighted ranking:
• weighted_score = (α × similarity) + (β × normalized_rating)

• ✨ Glassmorphic design
• 🎞️ IMDb quick-links
• 📝 Hoverable movie overviews
• 🔄 Smooth animations & transitions
• 📱 Fully responsive layout

┌────────────┐
│ Frontend │ ← Glassmorphic HTML UI
└─────┬──────┘
│ HTTP (JSON)
┌─────▼──────┐
│ Flask API │ ← app.py
└─────┬──────┘
│ Python Import
┌─────▼──────┐
│ ML Model │ ← model.py (TF-IDF + Logic)
└─────┬──────┘
│     |
┌─────▼──────┐
│ Dataset │ ← CSV
└────────────┘

✔ Clean separation of concerns
✔ One model, multiple recommendation paths

• Genres are normalized:
  • Science Fiction → Sci-Fi
• Converted to vectors using TF-IDF

• Computed using cosine similarity
• Implemented via linear_kernel (efficient on sparse matrices)

Instead of ranking only by similarity:

• weighted_score = (0.7 × genre_similarity) + (0.3 × normalized_rating)

This ensures:

• Highly similar movies are preferred
• Poorly rated movies don’t dominate results

The notebook includes meaningful, explainable plots:

• ⭐ Rating distribution (justifies mean imputation)
• 🎭 Genre frequency analysis
• 📊 Recommendation ranking (weighted score)
• ⚖️ Similarity vs Rating trade-off plots

These plots prove the model works — not just that it runs.

• POST /recommend/movie

{
  "movie_name": "The Avengers"
}

• Recommend by Genre

{
  "genres": ["Sci-Fi", "Thriller"]
}

git clone https://github.com/SACHIN-S-2004/Movie-Recommendation-System.git

cd Movie-Recommendation-System

pip install -r requirements.txt

python app.py

Open in browser:

http://127.0.0.1:5000