Movie Recommendation Using Unsupervised Method

Abstract

Everyone in our hectic world is occupied with their personal and professional lives. There is a maximum number of people who can watch movies in theaters or read books while passing the time. These days, online services like Netflix, Amazon, and others are extremely popular. A recommendation mechanism is necessary for this entire platform. I read some recently released publications to better grasp the current state of recommender systems as well as their potential future directions. The purpose of a movie recommendation system is to make recommendations for films based on the interests of various users. The customers would benefit from time savings when looking for popular movies of their preferred genre. It makes predictions about what movies a user will enjoy based on a data collection and takes into account the qualities of the movies they have already loved. Using a combination of two or more attributes, recommendation systems can suggest movies. Various elements, such as the movie's genre, directors, and performers, are taken into account when developing a movie recommendation algorithm. The cast, keywords, crew, and genres have been used to construct the recommendation system in this work. I start by gathering a data collection from many websites, including Kaggle. Then, using a few strategies, clean up and delete unneeded data. I employ machine learning algorithms like the count vector and cosine similarity algorithm after cleaning datasets. In our datasets, all algorithms perform excellently.