A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Interpretation of Dynamic Models Based on Neural Networks in the Form of Integral-Power Series
https://orcid.org/http://orcid.org/0000-0002-8816-0652
https://orcid.org/http://orcid.org/0000-0002-4082-8270
https://orcid.org/http://orcid.org/0000-0002-3764-9255
https://orcid.org/http://orcid.org/0000-0002-3256-5044
https://orcid.org/http://orcid.org/0000-0001-8353-7679
The paper is devoted to the problem of interpretation the dynamic models based on neural network. Proposed approach is conclude in the building of the interpretive model in the form of an analytical integral-power series, saving the dynamic and nonlinear properties of the primary neural network model. The purpose of the paper is development a method for interpreting dynamic models based on neural networks, providing high accuracy of the interpreting models construction. Scientific novelty consists in using of intego-power series in the form of multidimensional weight functions to build analytical models that interpret dynamic neural networks. This method allows to obtain an interpretation of dynamic neural network models while preserving their nonlinear and dynamic properties. Practical usefulness of the developed method consists in providing high accuracy and speed of interpreting models construction, allows to provide modeling of nonlinear dynamic states of objects in both test and functional modes of operation. The proposed method tested on the data of a test nonlinear dynamic object. The results of the experiment demonstrate high accuracy and speed of construction of analytical interpreting models.
Interpretation of Dynamic Models Based on Neural Networks in the Form of Integral-Power Series
https://orcid.org/http://orcid.org/0000-0002-8816-0652
https://orcid.org/http://orcid.org/0000-0002-4082-8270
https://orcid.org/http://orcid.org/0000-0002-3764-9255
https://orcid.org/http://orcid.org/0000-0002-3256-5044
https://orcid.org/http://orcid.org/0000-0001-8353-7679
The paper is devoted to the problem of interpretation the dynamic models based on neural network. Proposed approach is conclude in the building of the interpretive model in the form of an analytical integral-power series, saving the dynamic and nonlinear properties of the primary neural network model. The purpose of the paper is development a method for interpreting dynamic models based on neural networks, providing high accuracy of the interpreting models construction. Scientific novelty consists in using of intego-power series in the form of multidimensional weight functions to build analytical models that interpret dynamic neural networks. This method allows to obtain an interpretation of dynamic neural network models while preserving their nonlinear and dynamic properties. Practical usefulness of the developed method consists in providing high accuracy and speed of interpreting models construction, allows to provide modeling of nonlinear dynamic states of objects in both test and functional modes of operation. The proposed method tested on the data of a test nonlinear dynamic object. The results of the experiment demonstrate high accuracy and speed of construction of analytical interpreting models.
Interpretation of Dynamic Models Based on Neural Networks in the Form of Integral-Power Series
https://orcid.org/http://orcid.org/0000-0002-8816-0652
https://orcid.org/http://orcid.org/0000-0002-4082-8270
https://orcid.org/http://orcid.org/0000-0002-3764-9255
https://orcid.org/http://orcid.org/0000-0002-3256-5044
https://orcid.org/http://orcid.org/0000-0001-8353-7679
The paper is devoted to the problem of interpretation the dynamic models based on neural network. Proposed approach is conclude in the building of the interpretive model in the form of an analytical integral-power series, saving the dynamic and nonlinear properties of the primary neural network model. The purpose of the paper is development a method for interpreting dynamic models based on neural networks, providing high accuracy of the interpreting models construction. Scientific novelty consists in using of intego-power series in the form of multidimensional weight functions to build analytical models that interpret dynamic neural networks. This method allows to obtain an interpretation of dynamic neural network models while preserving their nonlinear and dynamic properties. Practical usefulness of the developed method consists in providing high accuracy and speed of interpreting models construction, allows to provide modeling of nonlinear dynamic states of objects in both test and functional modes of operation. The proposed method tested on the data of a test nonlinear dynamic object. The results of the experiment demonstrate high accuracy and speed of construction of analytical interpreting models.
Interpretation of Dynamic Models Based on Neural Networks in the Form of Integral-Power Series
Lect. Notes in Networks, Syst.
Arsenyeva, Olga (editor) / Romanova, Tatiana (editor) / Sukhonos, Maria (editor) / Tsegelnyk, Yevgen (editor) / Fomin, Oleksandr (author) / Polozhaenko, Sergii (author) / Krykun, Valentyn (author) / Orlov, Andrii (author) / Lys, Daria (author)
International Conference on Smart Technologies in Urban Engineering ; 2022 ; Kharkiv, Ukraine
2022-11-29
8 pages
Article/Chapter (Book)
Electronic Resource
English
| Online Contents | 1996
Representation of a multiple integral of special form by a series
| British Library Online Contents | 2013
Neural Networks Based Dynamic Load Modeling for Power System Reliability Assessment
| DOAJ | 2023