Balanced identification: testing hypotheses about the functioning of an object by experimental data processing

Alexander V. Sokolov

Keywords: balanced identification, data processing, optimization, testing hypotheses

TL;DR: formalization of the process of constructing mathematical models based on experimental data

Abstract:

The general formulation of the identification problem is considered. The issues of changing the set of feasible solutions when adding additional (changing existing) hypotheses are discussed. Balanced identification technology is used to connect mathematical models with data. Using real examples (plant physiology and pollution of the Kola Peninsula), the dynamics of some statistical estimates of modeling accuracy when adding “correct and incorrect” hypotheses is studied.

Submission Number: 59