Gene Expression Data Classification Using Artificial Neural Network Ensembles Based on Samples Filtering

Bioinformatics analysis based on microarray technology is facing serious challenges, due to the extremely high dimensionality of the gene expression data comparing to the typical small number of available samples. Single artificial neural network was unstable and inaccurate for classification. In this paper we introduce classifying gene expression data using artificial neural network ensembles based on samples filtering. Simulation tests were carried out to verify the proposed strategy using Leukemia data sets, and the test results were compared with those of single artificial neural network, bagging artificial neural network ensembles and support vector machine. The results indicated that our method is more stable and more accurate.