Output details
11 - Computer Science and Informatics
University of Exeter
Gene expression rule discovery and multi-objective ROC analysis using a neural-genetic hybrid
<28> This paper investigates the use of a multi-objective evolutionary algorithm and perceptron hybrid for the discovery of genes that can classify disease gene expression data. The method is used to discover a trade-off between maximal specificity and sensitivity and the minimum number of genes within the model for a number of datasets of varying difficulty. This is one of the first papers to use ROC analysis in this area and is an extended version of a conference paper presented at IEEE International Conference on Bioinformatics & Biomedicine (acceptance rate 17.2%). This journal has a 5-year impact factor of 0.64.