Extracting informative genes from unprocessed microarray data

Numerous feature selection methods have been developed to identify informative genes from a large pool of genes that are not involved in the array experiments. However, the integrity of the reported genes is still uncertain due to the applications of various pre-processing techniques to the microarray data by these methods and a lack of standard validation procedures to validate the significance of the genes. In this paper, we developed a feature extraction framework based on the hybrid genetic algorithm (GA) and neural network (ANN) to extract informative genes from the raw (unprocessed) microarray data. This approach has showed its efficacy in extracting informative genes for microarray data.