Applied problems of the identification and regression analysis in stochastic structures with multimodal properties

Nizhny Novgorod State Technical University n. a. R. E. Alekseev, Nizhny Novgorod, Russia

^a Corresponding author: vb.kulikov@yandex.ru

Abstract

We have been elaborating an approach founded on the identification of multimodal laws of the complex structure distribution in medicine, biology, chemistry of ultrapure materials and membrane technology as well as in technical applications. The method is based on the formulation and solution of inverse problems in mathematical physics for the respective probability density functions. The verification of the used algorithmic tools is carried out on model limited-scope samples. For stochastic structures and systems under study the method is supplemented with an original option of a regression analysis taking into account the identified stochastic laws displaying numerical parameters into the binary space. The proposed approach has been tested on clinical material in practical medicine.