On the properties of the classifier in case of small deviations of blood parameters from the norm

A mathematical model is described and implemented, intended for the numerical study of the ability of the statistical classification method to interpolate and extrapolate. The classifier developed by the authors is based on the polynomial-regression approach and has probabilistic estimates. It is used to assess the state of human health based on the parameters of laboratory analysis of peripheral blood. The blood base is considered with a small deviation from the norm.