Mutual Information Based ICA Applied to Microarray Data
Hoang, M. Thu, (Universite Rene Descartes), hoang@biomedicale.univ-paris5.fr,
Pham, Dinh Tuan, (Centre National de la Recherche Scientifique, IMAG), Dinh-Tuan.Pham@imag.fr, and
Thierry-Pascal Baum, (Centre National de la Recherche Scientifique, IMAG), Thierry-Pascal.Baum@imag.fr
Abstract
In this contribution, we demonstrate the use of the latest and performant algorithm for independent component analysis (ICA) developed by one of the author [1] for analyzing microarray data. We illustrate on T-ALL microarray data also discussed in [2] as an example in using gene profiling for prognosis. We intend to compare our method to similar statistical methods for relevance in extracting clinical knowledge from gene expression in human leukemia. [1] D. T. Pham (2004) Fast algorithms for Mutual Information Based Independent Component Analysis. IEEE Trans. on Signal Processing. To appear. (http://www-lmc.imag.fr/lmc-sms/Dinh-Tuan.Pham/BSS/AlgoICAinfof.ps.gz) [2] Ho\`ang T. and Parsons V. (2004) Bagging Survival Trees for Prognosis based on Gene Profiles, Compstat'2004. To appear. Physica Verlag.