PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

On Pairwise Kernels: An Efficient Alternative and Generalization Analysis
Hisashi Kashima, Satoshi Oyama, Yoshihiro Yamanishi and Koji Tsuda
In: 13th Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD), 27-30 Apr 2009, Thailand.

Abstract

Pairwise classification has many applications including network prediction, entity resolution, and collaborative filtering. The pairwise kernel has been proposed for those purposes by several research groups independently, and become successful in various fields. In this paper, we propose an efficient alternative which we call Cartesian kernel. While the existing pairwise kernel (which we refer to as Kronecker kernel) can be interpreted as the weighted adjacency matrix of the Kronecker product graph of two graphs, the Cartesian kernel can be interpreted as that of the Cartesian graph which is more sparse than the Kronecker product graph. Experimental results show the Cartesian kernel is much faster than the existing pairwise kernel, and at the same time, competitive with the existing pairwise kernel in predictive performance. We discuss the generalization bounds by the two pairwise kernels by using eigenvalue analysis of the kernel matrices.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Theory & Algorithms
ID Code:4406
Deposited By:Koji Tsuda
Deposited On:13 March 2009