A genetic algorithm for the classification of natural corks

摘要

In this paper, we explore the use of genetic algorithms (GA) for a classification problem encountered in wine industry: the classification of natural corks according to the defects of their heads. In particular, we are interested in the task of optimizing the parameters of an existing cork classification program. For this purpose, we introduce a GA-based approach that searches for good combinations from a huge search space. Experiments on both artificial and real data show the high effectiveness of this approach. This effectiveness justifies the use of this approach for daily operations in a real environment.