Learning Svm With Complex Multiple Kernels Evolved By Genetic Programming

  • Dioşan L
  • Rogozan A
  • Pecuchet J
  • 16

    Readers

    Mendeley users who have this article in their library.
  • 7

    Citations

    Citations of this article.

Abstract

Classic kernel-based classifiers use only a single
kernel, but the real-world applications have emphasised
the need to consider a combination of kernels, also
known as a multiple kernel (MK), in order to boost the
classification accuracy by adapting better to the
characteristics of the data. Our purpose is to
automatically design a complex multiple kernel by
evolutionary means. In order to achieve this purpose we
propose a hybrid model that combines a Genetic
Programming (GP) algorithm and a kernel-based Support
Vector Machine (SVM) classifier. In our model, each GP
chromosome is a tree that encodes the mathematical
expression of a multiple kernel. The evolutionary
search process of the optimal MK is guided by the
fitness function (or efficiency) of each possible MK.
The complex multiple kernels which are evolved in this
manner (eCMKs) are compared to several classic simple
kernels (SKs), to a convex linear multiple kernel
(cLMK) and to an evolutionary linear multiple kernel
(eLMK) on several real-world data sets from UCI
repository. The numerical experiments show that the SVM
involving the evolutionary complex multiple kernels
perform better than the classic simple kernels.
Moreover, on the considered data sets, the new multiple
kernels outperform both the cLMK and eLMK linear
multiple kernels. These results emphasise the fact that
the SVM algorithm requires a combination of kernels
more complex than a linear one in order to boost its
performance.

Author-supplied keywords

  • genetic programming
  • hybrid model
  • multiple kernel learning
  • svm

Get free article suggestions today

Mendeley saves you time finding and organizing research

Sign up here
Already have an account ?Sign in

Find this document

Authors

  • Laura Dioşan

  • Alexandrina Rogozan

  • Jean-Pierre Pecuchet

Cite this document

Choose a citation style from the tabs below

Save time finding and organizing research with Mendeley

Sign up for free