Modelling Hydroecomonitoring of Surface Water in Ukraine Using Machine Learning

Abstract

Providing the world's population with good-quality drinking water is one of the most important current global challenges. The authors suggest using artificial intelligence tools for factor analysis of water pollution of various origins, modelling probable parameter series, predicting the parameters of ongoing biological and chemical processes, and identifying data with a low level of reliability. The proposed method of analyzing environmental information allows to identify the factor influence of polluting compounds and to model the requested data series. The construction of the machine learning model described in the study involves selecting the most efficient information processing algorithm, adapting it to the training data set to build the required model design, further testing and calculating metrics. In order to improve the forecasting accuracy, a meta-classifier has been developed that combines several basic classifiers as part of an assembly model. Integration of the described methods into a hydroecological monitoring data processing system can increase the overall performance and correctness of information analysis, as well as make it possible to model scenarios of development of physical and chemical parameters.