http://www.diva-portal.org
This is the published version of a paper presented at 6th annual Big Data Conference at Linnaeus University, in Växjö, Sweden, 3-4 december, 2020.
Citation for the original published paper:
Daniel, W., Jusufi, I., Martins, R M., Kerren, A. (2020) Multiple Embeddings for Multivariate Network Analysis
In: 6th annual Big Data Conference at Linnaeus University, in Växjö, Sweden
N.B. When citing this work, cite the original published paper.
Permanent link to this version:
http://urn.kb.se/resolve?urn=urn:nbn:se:lnu:diva-99487
Multiple Embeddings for Multivariate Network Analysis
The visualization and visual analytics of large multivariate networks (MVN) continues to be a great challenge and will probably remain so for a foreseeable future. The field of Multivariate Network Embedding seeks to meet this challenge by providing MVN-specific embedding technologies that targets different properties such as network topology or attribute values for nodes or links. (Embeddings are relatively low-dimensional vector representations of the embedded items and they are well suited for similarity calculations.) Although many steps forward have been taken, the goal of efficiently embedding all aspects of a MVN remains distant. As a possible way forward we suggest a new angle of approach where, instead of trying to fit all aspects of a MVN into one embedding, the strategy would be to embed each property by itself and then find ways to combine these sets of embeddings.
*Contact: daniel.witschardr@lnu.se
http://cs.lnu.se/isovis/
QUESTION
Can embeddings of different types be combined to improve the quality of similarity calculations?
Conf ‘99, 1–2 Month 2099, City, Country
