| Author: | Lin, Da |
| Title: | Empowering graph neural network based recommendation system with few shot learning |
| Advisors: | Chung, Fu-lai Korris (COMP) |
| Degree: | M.Sc. |
| Year: | 2021 |
| Subject: | Recommender systems (Information filtering) Neural networks (Computer science) Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Computing |
| Pages: | iv, 47 pages : color illustrations |
| Language: | English |
| Abstract: | We are now living in an era characterized by information explosion. Hence, how to obtain information efficiently and quickly has become an urgent problem, and recommendation system is one representative solution. Traditional recommendation algorithms are based on sufficient data to make accurate recommendations, which often have poor performance in scenarios with sparse data. User cold start recommendation is very important in practice. Traditional methods solve problems more from the data level and obtain more auxiliary user information or item information. On the other hand, recently there are some methods that make use of meta learning models to solve the cold start problem, such as MetaHin, MELU, etc., but they are all based on the MAML model, which follows the idea of optimization-based meta learning. In this thesis, we propose an integrated graph neural network, metric based meta learning method to address user cold start problem. This is also the first time to use metric based method in user cold start problem. We validate our model on three data sets and compare it with two state-of-the-art cold-start recommendation models, namely, MetaHin and MELU. Our model has obtained the same performance as MELU, but is slightly weaker than MetaHin, while it has a significant performance improvement than the traditional models and the content-based cold-start models. Our approach and the experimental results obtained demonstrate the effectiveness of a new way of using GNN metric based meta learning method to solve the user cold start problem. |
| Rights: | All rights reserved |
| Access: | restricted access |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 5826.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 1.47 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/11388