Adaptive algorithm of classifcation on the missing data

Abstract

The problem of classi cation by data with gaps, bypassing the stage of their lling, is considered. An adaptive restructuring of algorithms is proposed as a result of the introduction of corresponding indicators into them. The indicators take into account the ow of current information, on the basis of which a decision is made to change the algorithm and the data processing technology itself at each cycle. Computational procedures are based on non-parametric estimation, are given their settings and the results of numerical modeling. Сборник проходит регистрацию в базе Scoupus.