The Graphlet Spectrum
Abstract
Current graph kernels suffer from two limitations: graph kernels based on counting particular types of subgraphs ignore the relative position of these subgraphs to each other, while graph kernels based on feature extraction by algebraic methods are limited to graphs without node labels. In this paper we present the graphlet spectrum, a system of graph invariants derived by means of group representation theory, that captures information about the number as well as the position of labeled subgraphs in a given graph. In our experimental evaluation the graphlet spectrum outperforms state-of-the-art graph kernels.
Cite
Text
Kondor et al. "The Graphlet Spectrum." International Conference on Machine Learning, 2009. doi:10.1145/1553374.1553443Markdown
[Kondor et al. "The Graphlet Spectrum." International Conference on Machine Learning, 2009.](https://mlanthology.org/icml/2009/kondor2009icml-graphlet/) doi:10.1145/1553374.1553443BibTeX
@inproceedings{kondor2009icml-graphlet,
title = {{The Graphlet Spectrum}},
author = {Kondor, Risi and Shervashidze, Nino and Borgwardt, Karsten M.},
booktitle = {International Conference on Machine Learning},
year = {2009},
pages = {529-536},
doi = {10.1145/1553374.1553443},
url = {https://mlanthology.org/icml/2009/kondor2009icml-graphlet/}
}