Graph neural network in recommender systems
The rapid increase in online data has made recommender systems crucial for managing information overload. These systems are highly valued for their ability to filter and recommend relevant content. However, one of the main issues for recommender systems is to learn accurate representations of user a...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/177135 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | The rapid increase in online data has made recommender systems crucial for managing information overload. These systems are highly valued for their ability to filter and recommend relevant content. However, one of the main issues for recommender systems is to learn accurate representations of user and items through their interaction and any other additional information. The reason why Graph Neural Networks (GNNs) have become popular in these systems is because they are well-suited to handle the graph-like nature of the data and are superior in learning graph representations. This project aims to cover an extensive review of the latest studies on GNN-based recommender systems. It categorizes these models by the kind of information they use and the types of recommendations they make. The project also reviews the obstacles encountered when applying GNNs to various datasets and examines how current research overcomes these difficulties. Additionally, it outlines emerging outlooks in the progression of this domain and compiles a list of significant research papers and their available open-source code. |
|---|