An investigation of the application of graph neural networks in recommendation systems

An investigation of the application of graph neural networks in recommendation systems

Matrix Factorization, popularized by the Netflix Prize, has established itself as the prevailing method for recommendation systems based on latent factor models. While traditional latent factor models like matrix factorization focus on capturing latent factors using linear algebra techniques, Graph...

Full description

Saved in:
Bibliographic Details
Main Author: Koh, Jaylene Jia Ying
Other Authors: Luu Anh Tuan
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2023
Subjects:
Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
Online Access:https://hdl.handle.net/10356/171973
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Matrix Factorization, popularized by the Netflix Prize, has established itself as the prevailing method for recommendation systems based on latent factor models. While traditional latent factor models like matrix factorization focus on capturing latent factors using linear algebra techniques, Graph Neural Networks extend this concept by considering more intricate relations within graphs. Hence, in this paper, we would explore the use of three types of state of the arts models: Graph Convolution Network, Graph Attention Network and GraphSAGE, to enhance a latent factor model by the graph structure and interactions.

Similar Items