Development of a Movie Recommendation System - MoviepleX

Abstract

The content recommendation model,  “Development of a Movie Recommendation System - MoviepleX” is aimed at providing accurate movie  recommendations to users, on the basis of similarity with  the movie they would enter for reference, using machine  learning algorithms, functions and metrics. It is built using  the tmdb_5000 dataset, taken from Kaggle. The data  consists of a number of features like cast, crew, genre,  budget, overview, runtime, tagline, popularity, production  unit and revenue corresponding to 4803 Hollywood  movies that are a part of the tmdb database. Recommendation engines are a subclass of information  filtering systems that seek to predict the 'rating' or  'preference' a user would give to an item, a movie in case  of a movie recommender. Streaming media services like  Netflix & Disney+ Hotstar employ highly efficient content  recommendation systems, which can play a huge role as  game-changers in a streaming service’s success or failure.  These content-based recommenders are what keep our  entertainment rhythm going, serving us the best material  out there, based on our own personal interests, choices,  likes & dislikes. Movie recommendation systems provide  a mechanism to assist viewers and subscribers of streaming  platforms by classifying movies based on similar interests  of users. A movie recommendation is important in our  social life due to its strength in providing enhanced  entertainment.  The model proposed in this paper uses machine learning’s  capability to identify patterns and build prediction and  recommendation mechanisms using provided data. A  machine learning web application was created for the  recommendation engine, which was deployed onto  Heroku, a container-based cloud Platform as a Service  (PaaS), used to deploy, manage, and scale modern apps.  The app deployment was made through Streamlit. By  having a webpage for the ML - application, it has been  made accessible and beneficial to public.