Non-sparse Multiple Kernel Learning for Fisher Discriminant Analysis

Authors:
Fei Yan;Josef Kittler;Krystian Mikolajczyk;Atif Tahir
Affiliations:
-;-;-;-
Venue:
ICDM '09 Proceedings of the 2009 Ninth IEEE International Conference on Data Mining
Year:
2009

Citing 0
Cited 5

Combining multiple kernels by augmenting the kernel matrix

MCS'10 Proceedings of the 9th international conference on Multiple Classifier Systems
Non-sparse multiple kernel fisher discriminant analysis

The Journal of Machine Learning Research
Multiple kernel learning with gaussianity measures

Neural Computation
Regularized multi-view learning machine based on response surface technique

Neurocomputing
A novel multiple Nyström-approximating kernel discriminant analysis

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

We consider the problem of learning a linear combination of pre-specified kernel matrices in the Fisher discriminant analysis setting. Existing methods for such a task impose an $\ell_1$ norm regularisation on the kernel weights, which produces sparse solution but may lead to loss of information. In this paper, we propose to use $\ell_2$ norm regularisation instead. The resulting learning problem is formulated as a semi-infinite program and can be solved efficiently. Through experiments on both synthetic data and a very challenging object recognition benchmark, the relative advantages of the proposed method and its $\ell_1$ counterpart are demonstrated, and insights are gained as to how the choice of regularisation norm should be made.