movie-recommender

This project demonstrates the use of basic algorithms in movie recommendation systems.

Requirements

Python 3.9.13

Steps to execute:

Run the following commands on command prompt terminal
(after python is installed)
Make sure to enter into the movie-recommender-master directory on command prompt(using cd) for the following steps

• Creating and entering into virtual environment
  • pip install virtualenv
  • python -m virtualenv myvenv
  • myvenv\Scripts\activate.bat
    
• Installing streamlit
  • pip install streamlit
    IMP: close the text editor and open it again (for streamlit installation to be effective)
• Execution
  On a new cmd terminal run
  • streamlit run app.py

How it works

• A movie is typed into the input box as shown in the output screenshot.
• On clicking Show Recommendation button, the selected movie and other suggested movies are displayed.
• Even if the typed movie does not exist/not in database, top matched movies are displayed
• Similarity index between movies is calculated (using KNN algorithm) based on cast, genre, director and keyword fields from the dataset

Algorithms used:

1. KNN algorithm
2. Quick sort algorithm
3. Search algorithm

Tech stack:

1. Python 3 for algorithms backend
2. Streamlit for frontend

Datasets:

1. trialdb- movie dataset containing various fields such as Title, genre, cast, release date etc.
2. posterdb- dataset containing movies titles and their poster image urls