Movie Recommendation System

About This Project:

This is a Streamlit web application that provides personalized movie recommendations based on user interests. The system employs a content-based filtering algorithm, which suggests movies with similar attributes to the ones the user has shown interest in. Below are some screenshots showcasing the app's interface:

Features:

This model uses a content-based filtering algorithm.

Content-Based Filtering:

• Content-based systems use item attributes (such as movie genres, cast, director, etc.) to make recommendations.
• Systems like Twitter and YouTube use similar algorithms to suggest content based on user preferences.
• The system forms embeddings (vectors) based on features, such as the movies you watch or the music you listen to.
• Recommendation Approach: It creates a vector of item features and compares them with the user’s preferences to recommend similar items.
• Advantages: Personalized suggestions based on past user behavior.
• Limitations: This may lead to excessive specialization (i.e., the system might only recommend items from the same categories, missing out on other potentially interesting items).

Data:

Data is sourced from the TMDb Movie Metadata Dataset.

Model Overview:

The model uses Cosine Similarity to compare movies:

1. Cosine Similarity is a metric that measures the similarity between two vectors.
2. Vectors are generated from movie features using a NumPy array.
3. The cosine_similarity() function calculates the similarity between two vectors, with values ranging from [0,1].
   • 0: Completely dissimilar.
   • 1: Completely similar.

Steps to Run the Project:

1. Clone the repository:

   https://github.com/Kyouma45/Movie_Recommendation_System

2. Create a conda environment after opening the repository make sure that you are using python>=3.7

   conda create -n movie python=3.7.10 -y
   

   conda activate movie
   

3. Install the requirements

   pip install -r requirements.txt
   

4. Generate models

   #run this notebook file to generate models
   Movie_Recommender_System.ipynb
   

5. Now run

   streamlit run app.py