water level forecast service
The "Flood" project aims to forecasting water levels for stationary hydrological posts with purpose of anti-flood measures.
Average forecast accuracy up to 95%
Brief Description of the Project
As part of the project, a machine learning library was developed,
as well as a recurrent neural network (RNN) with a redesigned mathematical learning apparatus
to improve forecasting accuracy. The developed neural network system "Flood" was tested during
the severe flood of 2021-2022 in the Republic of Bashkortostan and has the following functionality:
1. Neural network flood forecast for each hydrological post available in the database.
Currently, only the "Republic of Bashkortostan" region is available.
2. Neural network modeling of flood zones based on real and forecast data.
3. Obtaining historical and current data from open sources to populate the database in order
to predict water levels.
4. Imputation of missing data to improve the accuracy of predicted water levels.
Development Team
Evgeny Palchevsky
Project Manager, Senior Lecturer at the Department of Data Analysis and Machine Learning of the
Financial University under the Government of the Russian Federation.
Artem Kuzmichev
Project developer, student of the Department of Data Analysis and Machine Learning of the
Financial Student of the University under the Government of the Russian Federation. Field of
study: "Applied Computer Science".
Valery Koryakin
Project developer, student of the Department of Data Analysis and Machine Learning of the
Financial Student of the University under the Government of the Russian Federation. Field of
study: "Applied Computer Science".
Alexander Pyatunin
Project developer, student of the Department of Data Analysis and Machine Learning of the
Financial Student of the University under the Government of the Russian Federation. Field of
study: "Applied Computer Science".
Elina Soboleva
Project designer, student of the Department of Data Analysis and Machine Learning of the
Financial Student of the University under the Government of the Russian Federation. Field of
study: "Applied Computer Science".
Alexander Kamantsev
Project developer, student of the Department of Data Analysis and Machine Learning of the
Financial Student of the University under the Government of the Russian Federation. Field of
study: "Applied Computer Science".
Anton Pitsenko
Project developer, student of the Department of Data Analysis and Machine Learning of the
Financial Student of the University under the Government of the Russian Federation. Field of
study: "Applied Computer Science".