Models for predicting coffee yield from chemical characteristics of soil and leaves using machine learning

Abstract

BackgroundCoffee farming constitutes a substantial economic resource. This crop represents a source of income for several countries due to the high consumption of coffee drinks worldwide. Precise management of coffee crops involves collecting crop attributes (soil, plant), mapping, and applying inputs according to the plant's needs. This differentiated management is Precision Coffee Growing stands out for its increased yield and sustainability.ResultsThus, this research aimed to predict yield in coffee plantations by applying machine learning methodologies to soil and plant attributes. The data was obtained in a field of 54.6ha during two consecutive seasons, applying varied fertilization rates according to recommendations of soil attributes maps. Furthermore, monitoring leaf analysis maps seeks to establish a correlation between input parameters and yield prediction. The machine learning models obtained from this data efficiently predicted coffee yield. The best model demonstrated predictive fit results of 0.86 Pearson correlation. Soil chemical attributes did not interfere with the prediction models, indicating that this analysis can be dispensed with when applying these models.ConclusionThese findings have important implications for optimizing coffee management and cultivation, providing valuable insights for producers and researchers interested in maximizing yield using precision agriculture.This article is protected by copyright. All rights reserved.