COMPARISON OF LINEAR COLLECTIVES OF MULTIDIMENSIONAL NONPARAMETRIC REGRESSIONS
The technique of synthesis and the analysis of linear collectives of multidimensional nonparametric regressions is proposed. The building of collective components is based on the data of various volumes. Asymptotic properties of the collective are studied. The results of their comparison with the properties of the collective nonparametric regressions are analyzed with a uniform distribution of statistical data between its components.
Keywords: nonparametric regression, linear collective of decision functions, big samples, asymptotic properties, parallel computing