Content-Based Movie Recommendation System

This is a simple Movie Recommender System that suggests similar movies based on a user’s selection. It uses content-based filtering with cosine similarity over vectorized movie metadata (like genres, keywords, and descriptions). A basic web interface was created using Flask for user interaction.

---

Dataset

• Source: Movie metadata from TMDB (via Kaggle)
• The data was cleaned and preprocessed using Pandas.
• Features used for similarity:
  • genres
  • keywords
  • overview (description)

---

Libraries Used

• pandas, numpy – data manipulation and wrangling
• scikit-learn – cosine similarity & vectorization
• nltk – text preprocessing
• flask – backend for serving recommendations
• pickle – for saving/loading the similarity matrix

---

Recommendation Logic

• Extracted important text-based features.
• Combined them into a single string representation for each movie.
• Vectorized text using TfidfVectorizer.
• Computed cosine similarity between all movies.
• Given a movie, returned the top 5 most similar ones.

---

Web App

• Created a simple Flask app (app.py)
• Users can:
  • Select a movie
  • View recommended movies instantly
• Minimalistic and easy-to-use interface

---

Downloading the Model File

The similarity.pkl file is too large to include in the repo.

To run the recommender system locally, download the file using:

- pip install gdown
- python download_model.py

---

Author

• Ali Shayan – GitHub Profile

---

Feel free to fork the repo, suggest improvements, or report issues!