Fast Neighborhood Subgraph Pairwise Distance Kernel
Abstract
We introduce a novel graph kernel called the Neighborhood Subgraph Pairwise Distance Kernel. The kernel decomposes a graph into all pairs of neighborhood subgraphs of small radius at increasing distances. We show that using a fast graph invariant we obtain significant speed-ups in the Gram matrix computation. Finally, we test the novel kernel on a wide range of chemoinformatics tasks, from antiviral to anti carcinogenic to toxicological activity prediction, and observe competitive performance when compared against several recent graph kernel methods.
Cite
Text
Costa and De Grave. "Fast Neighborhood Subgraph Pairwise Distance Kernel." International Conference on Machine Learning, 2010.Markdown
[Costa and De Grave. "Fast Neighborhood Subgraph Pairwise Distance Kernel." International Conference on Machine Learning, 2010.](https://mlanthology.org/icml/2010/costa2010icml-fast/)BibTeX
@inproceedings{costa2010icml-fast,
title = {{Fast Neighborhood Subgraph Pairwise Distance Kernel}},
author = {Costa, Fabrizio and De Grave, Kurt},
booktitle = {International Conference on Machine Learning},
year = {2010},
pages = {255-262},
url = {https://mlanthology.org/icml/2010/costa2010icml-fast/}
}