movie-recommendation-system

Movie Recommendation System Models

Overview

This repository contains serialized model artifacts used by the Movie Recommendation System project.

The recommendation engine is a Content-Based Recommendation System built using Natural Language Processing techniques.

GitHub Repository: <YOUR_GITHUB_REPOSITORY_LINK>

---

Model Details

Technique

• TF-IDF Vectorization
• Cosine Similarity

Recommendation Type

Content-Based Filtering

Input

Movie title

Output

Top-N similar movies ranked by similarity score.

---

Files

movies.pkl

Processed movie metadata used by the recommendation engine.

similarity.pkl

Precomputed cosine similarity matrix used for fast recommendation retrieval.

---

Dataset

TMDB 5000 Movie Dataset

Source: https://www.kaggle.com/datasets/tmdb/tmdb-movie-metadata

---

Usage

Place the downloaded files inside:

models/

├── movies.pkl

└── similarity.pkl

Then load them in Python:

import pickle

movies = pickle.load(open("models/movies.pkl", "rb"))
similarity = pickle.load(open("models/similarity.pkl", "rb"))

---

Limitations

Current version uses movie overview text as the primary feature source.

Recommendations can be improved by incorporating:

• Genres
• Keywords
• Cast Information
• Director Information
• User Ratings

---

Intended Use

Educational and portfolio demonstration project showcasing:

• NLP
• Recommendation Systems
• TF-IDF Vectorization
• Similarity Search
• Machine Learning Workflows

---

License

This repository is intended for educational and non-commercial purposes.

license: ecl-2.0