Machine Learning Model Identification for Forecasting of Soya Crop Yields in Kazakhstan

In this article, using the example of soybean production in Kazakhstan, the features of using a new neuroprogramming method for analyzing data from field experiments and predicting yield are considered. It is shown that using historical statistics over several years, the program can create a trained model that is useful for predicting future values (profitability charts, anomalies, efficiency). The average error of the created neural yield model is 0.00894. The correlation coefficient of the developed neuromodel is 0.9602; determination coefficient - 0.9887. Based on the results of the work, a forecast of the yield of agricultural crops was obtained and recommendations were formulated to increase the yield of soybeans.