Movie-Recommendation-System

This project focuses on finding out the top five best picks similar to the recently watched movies. This is a content based recommendation system where the model suggest movies based on the watched content (like if someone is watching thriller then suggesting thriller movies).

Technology and Library Used

• Python
• Pandas
• NumPy
• Scikit-Learn
• NLTK
• Streamlit

Steps followed to build the project :

1. Dataset Downloaded from Kaggle , LINK : https://www.kaggle.com/datasets/tmdb/tmdb-movie-metadata and this contains two csv files tmdb_5000_movies.csv & tmdb_5000_credits.csv

2. Explored both the csv files and merged them based on the title column.

3. Done feature engineering where we kept only those features which will help us and removed all those which we generally not consider while recommending a movie. Then merged all such words which play an important role while recommending a movie into a single feature tags and removed others.

4. Perfomed some basic Natural Language Processing tasks like lowercasing and stemming. Then as computer can't deal with text so converted into vectors with the help of CountVectorizer function present in the scikit-learn.

5. Finally to find out the similarities between each vector, used cosine_similarity and according to that find out the five most similar movies.

6. Saved the movies list, similarity model into pickle file and load those file in the streamlit app which finally got deployed on HEROKU.

Project Link : https://movie-recommend-syt.herokuapp.com/

THANK YOU