🎬 Movie Recommendation System

A content-based movie recommendation system built from scratch using C++ and cosine similarity.

📌 Project Overview

This project demonstrates how to build a movie recommendation system using content-based filtering - one of the fundamental approaches in recommender systems. The system recommends movies based on genre similarity, using mathematical techniques like one-hot encoding and cosine similarity.

🧠 What is Content-Based Filtering?

Content-based filtering recommends items based on their features (in this case, genres). Here's how it works:

1. Feature Extraction: Convert movie genres into numerical vectors (one-hot encoding)
2. Similarity Calculation: Compare movies using cosine similarity
3. Ranking: Sort by similarity score and return top N recommendations

Simple Analogy

Think of movies as points in space. Movies with similar genres are "closer" together. The recommendation system finds the nearest movies to your selected movie!

📊 Cosine Similarity Explained Simply

Cosine similarity measures how similar two vectors are by calculating the cosine of the angle between them:

• 1.0 = Identical (perfect match)
• 0.0 = No similarity (completely different)
• -1.0 = Opposite (rare in this case)

Formula

Cosine Similarity = (A · B) / (||A|| × ||B||)

Where:

• A · B = Dot product of vectors
• ||A|| = Length of vector A
• ||B|| = Length of vector B

📁 Project Structure

MovieRecommender/
├── data/
│   └── movies.csv           # Movie dataset
├── include/
│   ├── Movie.h              # Movie class header
│   └── RecommenderSystem.h  # Recommender system header
├── src/
│   ├── main.cpp             # Main program & menu
│   ├── Movie.cpp            # Movie class implementation
│   └── RecommenderSystem.cpp # Core recommendation logic
├── Makefile                 # Build instructions
└── README.md                # This file

🚀 How to Build & Run

Prerequisites

• C++ compiler (g++ or clang++)
• Make (optional, but recommended)

Build Steps

bash
# Navigate to project directory
cd MovieRecommender

# Build the project
make

# Run the program
./movie_recommender

Or compile manually:

bash
g++ -std=c++17 -Iinclude -o movie_recommender src/main.cpp src/Movie.cpp src/RecommenderSystem.cpp
./movie_recommender

💡 How to Use

The system provides two ways to get recommendations:

Option 1: By Movie

Select a movie you like, and the system finds similar movies based on shared genres.

Option 2: By Genre

Enter your preferred genres, and the system recommends movies matching those genres.

Sample Dataset

The project includes 20 movies spanning genres like:

• Action, Adventure, Comedy, Crime, Drama
• Romance, Sci-Fi, Thriller, Animation, Fantasy

🎯 Features

• ✅ Dynamic genre extraction from CSV
• ✅ One-hot encoding for feature vectors
• ✅ Cosine similarity implementation (no external ML libraries)
• ✅ Interactive menu-driven interface
• ✅ Input validation and error handling
• ✅ Clean, modular OOP design
• ✅ Beginner-friendly with detailed comments

📈 Performance

Time Complexity

• Loading movies: O(M) where M = number of movies
• Building feature vectors: O(M × G) where G = number of genres
• Getting recommendations: O(M × G) per query

Space Complexity

• Movies storage: O(M × G)
• Genre index map: O(G)
• Feature vectors: O(M × G)

Scalability

• Works efficiently with 100+ movies
• For 1000+ movies: Consider indexing or caching similarity scores

🔧 Future Improvements

Want to take this further? Here are internship-level enhancements:

1. Weighted Genres: Give different weights to different genres
2. Rating Integration: Consider movie ratings in recommendations
3. Hybrid Scoring: Combine content-based and collaborative filtering
4. User Preferences: Save and learn from user feedback
5. Command-line Arguments: Allow passing CSV path as argument
6. Additional Features: Include actors, directors, year as features

🤝 Contributing

Feel free to fork this project and add your own improvements!