DESIGN OF NEURAL NETWORK FOR FORECAST ANALYSIS OF ELEMENTS-CONTAMINANTS DISTRIBUTION ON STUDIED TERRITORIES (ON EXAMPLE OF PAVLODAR CITY, KAZAKHSTAN)

Авторы

  • Ruslan Zairovich Safarov
  • Zhanat Kairollinovna Shomanova
  • Roza Zhumkenovna Mukanova
  • Yuri Gennadievich Nossenko
  • Alexandru Ilieș
  • Aleksander Konstantinovich Sviderskiy
  • Sarova Nurbanu

Ключевые слова:

neural network, contaminants, distribution, forecasting, modelling, GIS, environmental.

Аннотация

In the article we are presenting results of development appropriate method including neural network
for creating predictive map of elements-contaminants distribution (on example of Cr) on the territory of Pavlodar city
(Kazakhstan). Obtained method allows to get widen data. The data from 15 points were spread out into 500 points.
The average relative error at verification process was 9.45%. Architecture of well working model of neural network
is perceptron with one input neuron, which takes values of distances between given point and several nearest points,
10 hidden neurons, and 1 output neuron, which gives value of element concentration in specified point. Obtained
data were used in QGIS for creation of IDW interpolated map, which visualizes the information about concentration
distribution.

Загрузки

Опубликован

2019-12-05

Как цитировать

Ruslan Zairovich Safarov, Zhanat Kairollinovna Shomanova, Roza Zhumkenovna Mukanova, Yuri Gennadievich Nossenko, Alexandru Ilieș, Aleksander Konstantinovich Sviderskiy, & Sarova Nurbanu. (2019). DESIGN OF NEURAL NETWORK FOR FORECAST ANALYSIS OF ELEMENTS-CONTAMINANTS DISTRIBUTION ON STUDIED TERRITORIES (ON EXAMPLE OF PAVLODAR CITY, KAZAKHSTAN). SERIES CHEMISTRY AND TECHNOLOGY, (6), 86–98. извлечено от https://journals.nauka-nanrk.kz/chemistry-technology/article/view/1568

Выпуск

№ 6 (2019): Известия НАН РК. Серия химии и технологии

Раздел

Статьи