Authors: Piotr Artiemjew
Addresses: Department of Mathematics and Computer Sciences, University of Warmia and Mazury, Sloneczna 54, 10-710 Olsztyn, Poland
Abstract: This work extends the author's contribution to the Second International Conference of Soft Computing and Pattern Recognition (SocPar 2010) held at the University of Cergy Pontoise in December 2010. The current version is dedicated to the topic of gene separation algorithms and our best classification method based on weighted voting, which was investigated recently by Polkowski and Artiemjew. The DNA microarrays are a popular tool, useful to the research of gene expression. The exemplary application is, among others, to differentiate healthy and ill tissues, distinguishing some organisms' features or to check changes in gene expression by means of some additional factors. The huge amount of information obtained from DNA microarrays in the range of tens of thousands of genes causes many difficulties. Many algorithms, especially brute force methods, cannot be applied for this reason, and due to the low number of training objects tend to overfit. In this paper, we present two simple gene extraction methods compared by means of our best weighted voting classifier. The results of research show the high effectiveness of our approach, and full comparability with the results of the recent DNA microarray data mining competition.
Keywords: rough mereology; granular computing; rough sets; DNA microarrays; gene extraction; gene separation; classification; weighted voting; data mining.
International Journal of Data Mining, Modelling and Management, 2014 Vol.6 No.2, pp.110 - 126
Available online: 06 Jul 2014 *Full-text access for editors Access for subscribers Purchase this article Comment on this article