Advanced Soft-Computing techniques and Clustering Algorithm for Gene Expression Microarray Data Classification.
Olga Valenzuela, Fernando Rojas, Francisco Manuel Ortuño, José Luis Bernier, M. Jose Saez, Belen San-Roman, Luis Javier Herrera, Alberto Guillén, Ignacio Rojas · 2014
This paper is intended to meet two main objectives. The first is the development of an Parallel Genetic Algorithm using clustering fitness function, for Gene Expression Microarray automatic classification (is called PGA-GEM), which is focused on the processing data contained in genomic microarrays. Within this area, are of special interest the implemented fitness function and the genetic operators within evolutionary algorithm. The fitness function deter- mines the quality of the grouping obtained by statistical analysis of the data, discovering patterns that allow an automated classification and clustering of its. The second objective is to compare the results obtained by other algorithms with PGA-GEM , such as Support Vector Machines (SVM), showing than even it is not possible an superviser methodology, the presented algorithm can obtain similar results, performing an un-superviser training, than other well-known method as SVM