Genetic algorithm-neural network : feature extraction for bioinformatics data

摘要: With the advance of gene expression data in bioinformatics field, questions which frequently arise, for both computer and medical scientists, are genes significantly involved discriminating cancer classes significant with respect to a specific cancer pathology. Numerous computational analysis models have been developed identify informative from microarray data, however, integrity reported is still uncertain. This mainly due the misconception objectives study. Furthermore, application various preprocessing techniques has jeopardised quality data. As result, the integrity findings compromised by improper use techniques ill-conceived objectives research proposes an innovative hybridised model based on genetic algorithms (GAs) artificial neural networks (ANNs), extract highly differentially expressed for The proposed method can efficiently original set this has reduced variability errors incurred preprocessing techniques. novelty comes two perspectives. Firstly, emphasises extracting features high dimensional complex set, rather than improve classification results. Secondly, ANN compute fitness function GA rare context of feature extraction. Two benchmark taken prominent tumour development results show that respond different stages tumourigenesis (i.e. precision levels) may be useful early malignancy detection. extraction ability the proposed validated expected synthetic sets. In addition, bioassay used examine efficiency large, imbalanced multiple representation