The use of machine learning models in solving problems in the field of organic agriculture

Abstract

The article notes the active economic growth in the field of organic agriculture over the past 5 years. At the same time, the introduction of elements of artificial intelligence contributes to its more effective development. power with broad potential for ecological farming, development of the ability to form predictive models for selecting optimal sites for a given type of production using machine learning models. During the study, based on optimization methods, supervised neural network training models were built (set linear regression models, k-means method, computational model, random forest method and others). Under the quality metrics, the measured models are taken by the coefficient of determination R2 (i.e., the proportion of the variance of the dependent variable explained by the currents included in the model); model accuracy (for classifying models) and an alternative F1 score metric (accuracy and F1 score). The models analyzed the dependence of land use in organic farming such as an agricultural landscape; agrochemical soil conditions (primarily the presence of radionuclides, confirmed metals and herbicide/insecticide residues in the soil); remoteness from industrial agricultural complexes, cattle burial grounds, solid waste; indicators of isotherms and isobars, etc. were taken into account. The software solutions used were the Jupyter Notebook environment and the Google Colab cloud environment, as well as the standard libraries Pandas, NumPy, Scikit-learn, SciPy, Tensorflow, Matplotlib and others. Training and testing models were built based on a multiple linear regression model in a block of 70 to 30. The possibility of organizing organic agriculture on a specific land plot is turned off as a switching (dependent) variable. The resulting radius model evaluates the criteria for determining the dependence of a variable on an input one, and also gives a forecast of the possibility of transitioning a land plot to standard methods of organic farming..