An aligned subtree kernel for weighted graphs


In this paper, we develop a new entropic matching kernel for weighted graphs by aligning depth-based representations. We demonstrate that this kernel can be seen as an aligned subtree kernel that incorporates explicit subtree correspondences, and thus addresses the drawback of neglecting the relative locations between substructures that arises in the R-convolution kernels. Experiments on standard datasets demonstrate that our kernel can easily outperform state-of-the-art graph kernels in terms of classification accuracy.

Divisions: College of Engineering & Physical Sciences
College of Engineering & Physical Sciences > Systems analytics research institute (SARI)
Additional Information: © The authors
Event Title: 32nd International Conference on Machine Learning
Event Type: Other
Event Location: Lille Grand Palais
Event Dates: 2015-07-06 - 2015-07-11
Full Text Link: ... s/v37/bai15.pdf
Related URLs:
PURE Output Type: Conference contribution
Published Date: 2015
Authors: Bai, Lu
Rossi, Luca (ORCID Profile 0000-0002-6116-9761)
Zhang, Zhihong
Hancock, Edwin R.



Version: Published Version

Export / Share Citation


Additional statistics for this record