THE CONCEPT OF AN INFORMATION SYSTEM FOR FORECASTING THE TEMPERATURE REGIME OF THE EARTH’S SURFACE BASED ON MACHINE LEARNING

Abstract

The paper presents the concept of an information system for forecasting the temperature regime of the Earth’s surface using machine learning. Forecasting is based on historical data for a specific area. In order to increase the accuracy of forecasting results, an analysis of the features of climatic zones was carried out to identify patterns. A comparison of the dependence of the average monthly temperatures of the earth’s surface in countries depending on their location in climate zones was carried out.

The analysis of sources and scientific publications confirmed the relevance of the chosen research topic. Historical aspects of forecasting changes in climatic indicators are considered. Modern methods and approaches to temperature forecasting, their advantages and disadvantages are analyzed. An overview of the subject area was conducted and the regularities of temperature changes according to climate features were determined.

A comparison of temperature regimes for countries located in different climate zones was made. For clarity, graphs of temperature changes were plotted and average indicators were calculated for each climate zone.

The results of the study confirm the need to adjust the temperature forecast for certain areas, taking into account their location in a specific climate zone. The revealed regularities in the temperature regime of the countries indicate the need for an individual approach to forecasting and the use of such machine learning methods that are best adapted to the dependencies observed in the climate zone.

The architecture of the information system for forecasting future temperatures depending on the climatic features of the studied territories is proposed. A concept has been formed for further research to find more accurate and effective approaches to predicting climate parameters and achieving the goals of sustainable development.