Ivanchuk M. Mathematical modelling of expert medical prediction systems

The new method of the solution the classification problem is proposed. The method consists of two sets separation in the space using their e-nets building and separation. The concept of the set of possible values of epsilon is introduced. The properties of this set and the theorem of its convergence are proved. Weak convergence of the normalized difference of the empiric and theoretic separation curves to the normal distribution is proved. It makes possible to check the hypothesis of the place of theoretic separation curve in the concrete point. The algorithm of building e-nets for two sets is described. It is shown that for separating two sets one can use their e-nets, which considerably reduce the complexity of the separating algorithm.