Page 91 - Data Science Algorithms in a Week

In: Artificial Intelligence ISBN: 978-1-53612-677-8
Editors: L. Rabelo, S. Bhide and E. Gutierrez © 2018 Nova Science Publishers, Inc.

Chapter 4

EVOLUTIONARY OPTIMIZATION OF SUPPORT

VECTOR MACHINES USING GENETIC ALGORITHMS

*
Fred K. Gruber, PhD
Cambridge, Massachusetts, US

ABSTRACT

Support vector machines are popular approaches for creating classifiers in the
machine learning community. They have several advantages over other methods like
neural networks in areas like training speed, convergence, complexity control of the
classifier, as well as a more complete understanding of the underlying mathematical
foundations based on optimization and statistical learning theory. In this chapter we
explore the problem of model selection with support vector machines where we try to
discover the value of parameters to improve the generalization performance of the
algorithm. It is shown that genetic algorithms are effective in finding a good selection of
parameters for support vector machines. The proposed algorithm is tested on a dataset
representing individual models for electronic commerce.

Keywords: machine learning, support vector machines, genetic algorithms

INTRODUCTION

Support vector machines are popular approaches for developing classifiers that offer
several advantages over other methods like neural networks in terms of training speed,

* Corresponding Author Email: fgruber@ieee.org.