Feature Selection and Discretization for Microarray and Other Biological Data
Ferreira, A.
;
Figueiredo, M. A. T.
Feature Selection and Discretization for Microarray and Other Biological Data, Proc Inforum - Simpósio de Informática, Coimbra, Portugal, Vol. 1, pp. 60 - 65, September, 2011.
Digital Object Identifier:
Download Full text PDF ( 154 KBs)
Abstract
Tumor and cancer detection from microarray data are important bioinformatics problem. Since microarray datasets typically have a very large number of features and small number of instances, learning algorithms are thus confronted with the curse of dimensionality, and need to address it in order to be
effective. Moreover, it is common to have multi-class problems on microarray data, in which existing methods tend to perform worse than on binary classification problems. This paper proposes unsupervised feature discretization and selection methods suited for microarray data. The experimental results show that the
proposed techniques yield results comparable or better than previous approaches.