Gene selection and parameter determination of support vector machines based on BPSO algorithm
by Shutao Li, Xixian Wu, Mingkui Tan
International Journal of Modelling, Identification and Control (IJMIC), Vol. 8, No. 4, 2009

Abstract: The advent of microarray technology allows researchers to perform cancer diagnosis on gene level. However, the microarray data usually contains a large quantity of irrelevant, noisy and redundant genes which may seriously deteriorate the prediction accuracy. What's more, microarray data always suffers from the problem of a relatively small number of samples, which raises the difficulties in diagnosis. Accordingly, gene selection plays an important role in cancer diagnosis using microarray data. As a state-of-the-art classifier, support vector machine (SVM) has been successfully applied in solving the small-sample recognition problems and therefore is very suitable for microarray data processing. However, to improve the recognition accuracy, it is essential to choose certain good enough parameters for SVM in applications. This study proposes a new gene selection method based on a binary version of particle swarm optimisation (BPSO) algorithm and SVM. In the method, the task of gene selection and parameter tuning of SVM is performed simultaneously by BPSO. To verify the proposed method, three benchmark dataset (leukaemia, breast and colon) are studied. The experimental results show that the proposed method is highly effective and competitive compared with other methods.

Online publication date: Wed, 09-Dec-2009

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Modelling, Identification and Control (IJMIC):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com