Evolutionary weighting of image features for diagnosing of CNS tumors

This paper concerns an application of evolutionary feature weighting for di agnosis support in neuropathology. The original data in the classification task are the microscopic images of ten classes of central nervous system (C NS) neuroepithelial tumors. These images are segmented and described by the features characterizing regions resulting from the segmentation process. T he final features are in part irrelevant. Thus, we employ an evolutionary a lgorithm to reduce the number of irrelevant attributes, using the predictiv e accuracy of a classifier ('wrapper' approach) as an individual's fitness measure. The novelty of our approach consists in the application of evoluti onary algorithm for feature weighting, not only for feature selection. The weights obtained give quantitative information about the relative importanc e of the features. The results of computational experiments show a signific ant improvement of predictive accuracy of the evolutionarily found feature sets with respect to the original feature set. (C) 2000 Elsevier Science B. V. All rights reserved.