Please use this identifier to cite or link to this item:
https://dspace.ncfu.ru/handle/20.500.12258/18497
Title: | A method for predicting the evolvement of an emergency zone based on artificial neural networks |
Authors: | Petrenko, V. I. Петренко, В. И. Tebueva, F. B. Тебуева, Ф. Б. Antonov, V. O. Антонов, В. О. Svistunov, N. Y. Свистунов, Н. Ю. Kabinyakov, M. Y. Кабиняков, М. Ю. Garanzha, A. V. Гаранжа, А. В. Zavolokina, U. V. Заволокина, У. В. |
Keywords: | Neural networks;Artificial neural networks;Predicting the evolvement of an emergency zone |
Issue Date: | 2021 |
Publisher: | American Institute of Physics Inc. |
Citation: | Petrenko, V. I., Tebueva, F. B., Antonov, V. O., Svistunov N. Yu., Kabinyakov M. Yu., Garanzha, A. V., Zavolokina, U. V. A method for predicting the evolvement of an emergency zone based on artificial neural networks // AIP Conference Proceedings. - 2021. - Том 2402. - Номер статьи 050044. - DOI10.1063/5.0074027 |
Series/Report no.: | AIP Conference Proceedings |
Abstract: | This article describes a method for evolvement predicting of an emergency zone on the example of a forest fire using neural networks and applying the Hurst exponent to increase the accuracy and reduce the computational complexity of the algorithm. The inputs in this method are a set of environment state maps and the fire characteristics. The lack of a sufficient number of maps and a large amount of available data has a negative effect on the forecast accuracy and the overall system performance. It proposed to divide the fire area into sectors and perform the forecast by sectors with an accuracy that depends on the sector behavior characteristics. The sector behavior characteristic in this paper presented by the persistence analysis of the sector vectors, and be replaced by any other indicator. Persistence indicates trend stability in the behavior of the time series and, accordingly, the forecast reliability. In the case of low forecast reliability, the sector is iteratively divided into segments, which are analyzed and characterized. In this case, forecasting carried out for each sector and segment, which generally increases overall forecast accuracy. The paper presents the key mathematical calculations and the results of experimental studies. |
URI: | http://hdl.handle.net/20.500.12258/18497 |
Appears in Collections: | Статьи, проиндексированные в SCOPUS, WOS |
Files in This Item:
File | Size | Format | |
---|---|---|---|
scopusresults 1971 .pdf Restricted Access | 64.73 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.