Harvest prediction model based on public data for large regions

Abstract

This article describe harvest prediction model for the country or for the big region on the public available data. In the article are analysed impact of main fertilizers component and environmental variables to the grain harvest The aim of the article was to create regression model, which best describes grain harvest prediction on public (free) available data. Created final regression model explain 78% (R2) of the variation in the harvest result. Presented model show, that prediction accuracy significantly increase if environmental variables are added. Prediction accuracy (RMSE) of the final regression model was 3,89. All calculation was made on the example of the Germany.