Gubarev V. F., Miliavskyi Yu. L.
Features of modeling and identification of cognitive maps under uncertainty

A process of complex systems identification is examined. It is established that it is impossible to create a universal identification method. Only for a well-identifiable system with a high signal-to-noise ratio for each individual system mode, a high-quality model can be reconstructed. In other cases, if modes with sufficiently small signal-to-noise ratio exist, only a surrogate model can be obtained. For cognitive maps, theoretical foundations are developed, which may be used in approaches to find a surrogate model and then to improve the result using different tuning and learning algorithms. Numerical simulation was used to analyze the identification process. © 2023, Springer Science+Business Media, LLC, part of Springer Nature.

Keywords: cognitive map, complex system, ill-conditioning, regularization, subspace method, system identification, Cognitive systems, Identification (control systems), Signal to noise ratio, Uncertainty analysis, Cognitive maps, High signal-to-noise ratio, Identification method, Ill-conditioning, Modelling and identifications, Regularisation, Subspace method, Surrogate modeling, System-identification, Uncertainty, Large scale systems

Cite:
Gubarev, Miliavskyi. (2023). Features of modeling and identification of cognitive maps under uncertainty. Cybernetics and Systems Analysis, 59 (4), 43–59. doi: https://doi.org/10.1007/s10559-023-00590-2 http://jnas.nbuv.gov.ua/article/UJRN-0001415723