Artificial intelligence

Abstract: The paper proposes a new evolutionary approach to classifier ensemble design. The proposed approach is developed on the basis of genetic algorithm with modified realization scheme as applied to the optimization of feature set decomposition into the subsets, which define the individual ensemble’s classifiers and provide the high classification accuracy. During optimization both individual classifiers’ parameters and the ensemble parameters are defined. With the approach a few ensembles were designed for several datasets from machine learning database and for one real medical dataset. The comparative testing shows the advantages of the proposed approach for multivariate data analysis with great number of features.

Keywords:

References: