MOVIE RECOMMENDATION SYSTEM USING DEEP LEARNING.
Loading...
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Malaysia Sarawak (UNIMAS)
Abstract
Description
The ubiquitous demands of personalized content have driven recommender systems out of the traditional algorithmic paradigm. This report explores the current industry issues, including data sparsity and the failure to capture complex preferences, by building a movie recommendation system using modern deep-learning methods. The models used are Neural Collaborative Filtering (NCF), Deep Autoencoders (DAE), and Singular Value Decomposition (SVD), and each of them is designed to capture complex user-item interactions. Hit Rate @ 10 (HR@10) and Normalized Discounted Cumulative Gain @ 10 (NDCG@10) were used to make a complete evaluation. The research ends with the implementation of a web-based application that combines the best model, thus providing accurate recommendations and increasing the overall user satisfaction.