I.M. Antonyan, Y.V. Roshchin, O.I. Zelenskyi, F.G. Moshel, T.A. Nalbandyan, A.Y. Sokolov, V.A. Goryachaya, Y.M. Ugryumova

New opportunities for diagnosis and monitoring of patients with prostate conditions

The method of estimation of informativeness (importance) variables of diagnostic system models which are derived from the learning machine theory of artificial neural networks (ANN) was developed. Comparison of the quality of approximation of data using linear (linear multiple regression) and nonlinear (in the form of unidirectional and radial-basic trainees ANN) models were conducted. Evaluation of informativeness of controlled state variable of biomedical system elements (MBS) were obtained based on the accuracy of their measurements using linear and nonlinear diagnostic models.