A NEAT-based multiclass classification method with class binarization
Zhenyu Gao, Gongjin Lan · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2021
Multiclass classification is a fundamental and challenging task in machine learning. Class binarization is a popular method to achieve multiclass classification by converting multiclass classification to multiple binary classifications. NeuroEvolution, such as NeuroEvolution of Augmenting Topologies (NEAT), is broadly used to generate Artificial Neural Networks by applying evolutionary algorithms. In this paper, we propose a new method, ECOC-NEAT, which applies Error-Correcting Output Codes (ECOC) to improve the multiclass classification of NEAT. The experimental results illustrate that ECOC-NEAT with a considerable number of binary classifiers is highly likely to perform well. ECOC-NEAT also shows significant benefits in a flexible number of binary classifiers and strong robustness against errors.