Learning on Graphs: Supervised and Unsupervised Methods
Public Deposited- Last Modified
- May 14, 2019
- Creator
-
Emmons, Scott
- Affiliation: College of Arts and Sciences, Department of Mathematics
- Abstract
- We study two methods for learning from network graph data. First, we present a novel method for the unsupervised learning problem of community detection. The proposed method is, to the best of our knowledge, the first enabling users to "zoom in" and "zoom out" on communities with varying levels of focus on network metadata. Second, we review Decagon, a system proposed by Zitnik et al. for the supervised learning task of link prediction. On a biomedical benchmark dataset, Decagon achieves state-of-the-art prediction accuracy. This work adds to the network scientist's machine learning toolkit, illustrating its power in a biomedical domain with significant public health impact.
- Date of publication
- April 2019
- Keyword
- DOI
- Resource type
- Note
- Funding: James S. McDonnell Foundation 21st Century Science Initiative - Complex Systems Scholar Award grant #220020315
- Advisor
- Mucha, Peter
- Degree
- Bachelor of Science
- Academic concentration
- Mathematics
- Honors level
- Honors
- Degree granting institution
- University of North Carolina at Chapel Hill
- Graduation year
- 2019
- Language
- English
- Date uploaded
- April 26, 2019
Relations
- Parents:
- In Collection:
This work has no parents.
Items
| Thumbnail | Title | Date Uploaded | Visibility | Actions |
|---|---|---|---|---|
|
|
2019_Emmons.pdf | April 26, 2019 | Public | Download |