Movie Recommendation Models Using Machine Learning
Shourya Chawla, Sumita Gupta, Rana Majumdar · 2021 5th International Conference on Information Systems and Computer Networks (ISCON) · 2021
A recommendation system is essentially a kind of data filtering. Recommendation systems are very important in today's world since they give users suggestions based on their interests and needs when they are looking for something. Different types of recommendation systems are utilized on numerous platforms and have become an integral feature of a variety of applications. Movie recommendation systems aim to assist movie lovers by recommending what movie to watch without requiring them to go through the time-consuming and complex process of selecting from a vast number of movies ranging from thousands to millions. The goal of this paper is to propose a hybrid movie recommendation system based on the MovieLens dataset. The aim is to combine the user's personalization with the movie's overall features, such as genre, popularity, and so on. A popularity model, a content-based model, a collaborative filtering model, and a model based on latent factors have been used. Each model's hyperparameter tuning, testing accuracy, and evaluation are meticulously carried out. A combined linear model is created by combining the predictions of latent factor based and collaborative filtering methods, which exhibits an improvement in the accuracy of anticipated ratings. To address the constraints of individual models, all the methods are combined to build a hybrid model to get the final suggestion list for users. In terms of quality and diversity of recommendations, the hybrid model performs better than the separate models. The study contains a thorough assessment of each model.