This project aims to enhance the movie-watching experience by developing a content-based recommendation engine leveraging cosine similarity. The primary objective is to provide users with personalized movie suggestions based on the content of movies they enjoy. As a content-based recommendation system, it effectively addresses the cold start problem.
The dataset used for this project is the TMDB 5000 dataset.
Approach:
- Data Preprocessing: This includes data cleaning, stemming, and feature engineering.
- Text Vectorization Technique: Utilizing cosine similarity vectorization to convert movie descriptions into numerical vectors.
- Displaying the top 3 closest vectors to the user's input.
- Utilizing the IMDB website to retrieve images based on movie names.
This project represents a valuable contribution to enhancing user experiences in the realm of movie recommendations while effectively mitigating the cold start problem.
Here's a demonstration!!