Learning from graphs with structural variation
Machine Learning
2018-07-02 v1 Data Structures and Algorithms
Machine Learning
Abstract
We study the effect of structural variation in graph data on the predictive performance of graph kernels. To this end, we introduce a novel, noise-robust adaptation of the GraphHopper kernel and validate it on benchmark data, obtaining modestly improved predictive performance on a range of datasets. Next, we investigate the performance of the state-of-the-art Weisfeiler-Lehman graph kernel under increasing synthetic structural errors and find that the effect of introducing errors depends strongly on the dataset.
Keywords
Cite
@article{arxiv.1806.11377,
title = {Learning from graphs with structural variation},
author = {Rune Kok Nielsen and Andreas Nugaard Holm and Aasa Feragen},
journal= {arXiv preprint arXiv:1806.11377},
year = {2018}
}
Comments
Presented at the NIPS 2017 workshop "Learning on Distributions, Functions, Graphs and Groups"