TY - GEN

T1 - Curve Fitting for Exponential Polynomials from Interval Data

AU - Zvyagin, Maxim A.

AU - Shary, Sergey P.

N1 - Публикация для корректировки.

PY - 2023

Y1 - 2023

N2 - The article is devoted to solving the curve fitting problem for exponential polynomials from interval data. The goal of the work is to describe, implement and test the algorithm of curve fitting for exponential polynomials. The curve fitting algorithm is a generalization of the Maximal Compatibility Method for a given class of functions. The key idea is to implement a measure of data compatibility and to find the maximum of this measure. Thus, the problem of curve fitting boils down to the problem of a non-smooth non-concave conditional maximization of a function called recognizing functional. Subgradient methods, the penalty function method and multistart technology were used to solve this problem. In the course of the work, an explicit form of optimization problem was built, a software implementation was made, several examples demonstrating the performance of technology were built. The results of the work can be used in applied research to solve the problem of curve fitting, where the uncertainty of the data can be described with the help of intervals and the guaranteed nature of the estimates is required.

AB - The article is devoted to solving the curve fitting problem for exponential polynomials from interval data. The goal of the work is to describe, implement and test the algorithm of curve fitting for exponential polynomials. The curve fitting algorithm is a generalization of the Maximal Compatibility Method for a given class of functions. The key idea is to implement a measure of data compatibility and to find the maximum of this measure. Thus, the problem of curve fitting boils down to the problem of a non-smooth non-concave conditional maximization of a function called recognizing functional. Subgradient methods, the penalty function method and multistart technology were used to solve this problem. In the course of the work, an explicit form of optimization problem was built, a software implementation was made, several examples demonstrating the performance of technology were built. The results of the work can be used in applied research to solve the problem of curve fitting, where the uncertainty of the data can be described with the help of intervals and the guaranteed nature of the estimates is required.

UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-85171981111&origin=inward&txGid=87ee9b1715be4fe9ea04b5f09f5a6c2f

UR - https://www.mendeley.com/catalogue/8fd6bd58-b44d-33b7-a7d6-24342a94ef3a/

U2 - 10.1109/edm58354.2023.10225231

DO - 10.1109/edm58354.2023.10225231

M3 - Conference contribution

SN - 9798350336870

SP - 1880

EP - 1883

BT - 24th IEEE International Conference of Young Professionals in Electron Devices and Materials, EDM 2023; Novosibirsk; Russian Federation; 29 June 2023 до 3 July 2023

PB - Institute of Electrical and Electronics Engineers (IEEE)

ER -