A set of data on retrospective grain yield for neural network modeling

Abstract

Abstract Theoretical, methodological and applied aspects of creating a database (DB) of long-term retrospective information about the yield of grain crops are considered. The database is designed for statistical processing and modeling of multi-year data. Statistical information on annual levels of grain yield is the basis for planning, forecasting, management and optimization of agricultural production. To implement these tasks and their information support, we use design methods and models that allow us to build relational databases that provide statistical and neural network modeling of interannual variability of yield levels. The data storage format (*.csv) is justified, which provides in-depth processing and statistical analysis using built-in Python libraries. The review statistical analysis of productivity on the example of grain crops grouped by annual intervals is presented, and their features are revealed.