Alena A. Zakharova, Aleksandr G. Podvesovskii, Viktoria P. Tolstenok

Bryansk State Technical University

The article discusses the features of biomedical data and their influence on the choice of classifiers in the development of the ensemble of models. All biomedical data are conventionally divided into five groups, so they can have a different nature, and scaling is necessary for their systematization. The article also presents several features of biomedical data. Biomedical features have a direct impact on the choice of ensemble formation method (depending on the desired results), as well as on the choice of classifiers included in the ensemble. The following classifiers are presented in the article: correlation analysis, associative rules, Kohonen maps and decision trees. Thus, the features of the input data during the formation of the ensemble must be considered at the initial stage in order to avoid further errors when working on the ensemble and interpretation of the results.

biomedical data, features of biomedical data, ensemble of models, data mining