PECULIARITIES OF USING METHODS OF INTELLECTUAL DATA ANALYSIS TO SOLVE THE PROBLEMS OF ENVIRONMENTAL SAFETY OF ATMOSPHERIC AIR

Abstract

The process of environmental safety management in the context of reducing negative impacts on the environment is described. A generalized structural model of the environmental safety management process based on methods and technologies of intellectual analysis of monitoring data is proposed. Possibilities of adaptation and improvement of a number of the most known algorithms of data mining, such as C4.5, K-means, method of reference vectors (SVM), kNN, naive Bayesian classifier, Apriori algorithm, for data analysis of atmospheric air monitoring network are investigated. On the example of data on concentrations of pollutants in the air of Kryvyi Rih city (Dnipropetrovsk region) diagrams are constructed: scattering of concentrations of dust and nitrogen dioxide; dissipation of air temperature and concentration of sulfur dioxide. Examples of practical use of separate methods for the purpose of dangerous situations detection are resulted. The aim of the article is to analyze the principles and methods of environmental safety management based on the intellectual analysis of atmospheric air monitoring network data. Scientific novelty. A conceptual model of ecological safety management of urban areas based on ecological monitoring data is proposed, which differs from its analogues by new possibilities based on the use of modern information technologies of data mining. The conclusions emphasize that the authors of the publication identify prospects for the application of methods and tools of data mining for information support of decision-making aimed at assessing the effects of man-made impact and reducing the burden on the environment. A number of methods and algorithms are proposed, which make it possible to estimate the values of unknown characteristics and parameters according to known data, and examples of their use are shown.