N V.V. THE METHOD OF ESTIMATION OF THE ECOSYSTEMS STABILITY LEVEL BY USE OF AUTOMATIC LEARNING

Abstract

This work deals with the automation of ecosystem stability level assessment. Until now carrying out such assessment has been available only to high-class experts. That has complicated broad use of this estimation. However application of machine learning for processing of initial information on the basis of the decision-making logic permits to automate this procedure and make it quicker and more available for a wide range of experts. The article presents the practical description of using a technology of automation of forecasts making it possible to determine risks for the population residence in a particular area. The article also presents a constructed model of ecosystems stability in Russia. Model accuracy was more than 80%. The best algorithm of the data processing and its best parameters has been found for the computer assessment of the ecosystem stability level. It allows automatization of the complicated process of the expert area classification. This technology can be applied at the pre-design stage for any construction and included into the information system of the enterprise in terms of the decision support mechanism.

Keywords

• моделирование
• устойчивость экосистем
• машинное обучение
• автоматизация
• врожденная заболеваемость