Please use this identifier to cite or link to this item: https://sci.ldubgd.edu.ua/jspui/handle/123456789/18393
Title: Identification of Dynamic Regime Transition Models Between Erlang and Exponential Distributions
Authors: Tuluchenko, H.Ya.
Soviak, I.M.
Malanchuk, M.I.
Keywords: dynamic mixture model
Erlang and exponential distributions
Issue Date: Jun-2026
Publisher: International Conference of Young Mathematicians The Institute of Mathematics of the National Academy of Sciences of Ukraine
Citation: Tuluchenko H. Ya., Soviak I. M., Malanchuk M. I. Identification of Dynamic Regime Transition Models Between Erlang and Exponential Distributions. In: International Conference of Young Mathematicians, The Institute of Mathematics of the National Academy of Sciences of Ukraine, Kyiv, Ukraine, June 3–5, 2026.
Series/Report no.: International Conference of Young Mathematicians (ICYM), Proceedings of the Conference, 2026;
Abstract: A dynamic model is proposed for transition between Erlang and exponential distributions using a sigmoid weighting function. The model captures Erlang behavior at small values and exponential tails at large values. Parameters are estimated via maximum likelihood, and numerical results on synthetic data show stable estimation and improved fit compared to classical models.
Description: The study presents a dynamic statistical model describing a smooth transition between Erlang and exponential distributions using a sigmoid weighting function. The model enables flexible representation of data with different distributional regimes and heavy-tail behavior. Parameter estimation is performed via the maximum likelihood method, and numerical experiments on synthetic data confirm the effectiveness and improved fitting performance of the proposed approach compared to classical mixture models.
URI: https://sci.ldubgd.edu.ua/jspui/handle/123456789/18393
Appears in Collections:2026

Files in This Item:
File Description SizeFormat 
Identification of Dynamic Regime Transition Models.pdf170.05 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.