Abstract:
The problem of determining the ability to extrapolate a statistical classifier intended for assessing the state of human health by the parameters of peripheral blood is considered. A numerical method is used to study the characteristics of the set obtained from the training in the process of gradually increasing distortion. The extrapolation properties of the classifier and the dynamics of the probabilistic estimates generated by it are described.
Keywords:human health state, body system, peripheral blood, classification, polynomial regression, learning set.
Fulltext:
PDF file (5353 kB)