International Conference on Machine Learning, Date: 2010/06/21 - 2010/06/24, Location: Haifa, Israel
Publication date:
2010-01-01
Pages:
255 -
262
Publisher:
Omnipress; Madison, WI, USA
Proceedings of the 26th International Conference on Machine Learning
Author:
Costa, Fabrizio
De Grave, Kurt
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 anticarcinogenic to toxicological activity prediction, and observe competitive performance when compared against several recent graph kernel methods.