PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Kernelising the Ihara zeta function
Furqan Aziz, Richard Wilson and Edwin Hancock
Computer Analysis of Images and Patterns Volume 6854, pp. 219-227, 2011. ISSN 978-3-642-23671-6

Abstract

The Ihara zeta function, related to the number of prime cycles in a graph, is a powerful tool for graph clustering and characterization. In this paper we explore how to use the Ihara zeta function to define graph kernels. We propose to use the coefficients of reciprocal of Ihara zeta function for defining a kernel. The proposed kernel is then applied to graph clustering.

EPrint Type:Article
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Machine Vision
ID Code:8494
Deposited By:Edwin Hancock
Deposited On:03 February 2012