Forecasting of agricultural production volumes using methods of data mining

Abstract

In this article, the future values of indicators were forecasted for production of grains and legumes on farms in Cherkasy region based on the time series expressed in physical units. Time series analysis as one of the data mining techniques was used during the research in order to make a forecast of production using the data (based on the model of dynamic series) from past years to predict the future production volumes. This method contains the following steps: a graphical analysis (allows you to choose the model equation in the best way), separation and analysis of deterministic components of the series, smoothing and filtering of time series, study of random components, construction and testing for the adequacy of the time series model, forecasting the behavior of the time series based on the conducted research.