Journals Proceedings

Abstract

Thanks to genetic expression, biomarkers related to much kind of diseases could be obtained. However, depending on the previous experimental constraints, these biomarkers can become too many, especially if they are intended to be used in a classification process. Multi-Objective Genetic Algorithm (MOGA) has been thought of as a suitable solution to get a tradeoff among candidate biomarkers and classification accuracy. In this sense, genetic expression and Non Sorting Genetic Algorithm (NSGA) are combined in order to obtain the minimum number of possible genes that achieve the maximum classification accuracy. Only with a few genes, the classifier is expected to be more computationally efficient and faster than with all the genes, besides not endangering the final accuracy.