Comparison Between Machine Learning Models for Yield Forecast in Cocoa Crops in Santander, Colombia

The identification of influencing factors in crop yield (kg·ha-1) provides essential information for decision-making processes related to the prediction and improvement of productivity, which gives farmers the opportunity to increase their income. The current study investigates the application of mu...

Full description

Bibliographic Details
Main Authors: Lamos-Díaz, Henry, Puentes-Garzón, David Esteban, Zarate-Caicedo, Diego Alejandro
Format: Online
Language:eng
spa
Published: Universidad Pedagógica y Tecnológica de Colombia 2020
Subjects:
Online Access:https://revistas.uptc.edu.co/index.php/ingenieria/article/view/10853
_version_ 1801706087205830656
author Lamos-Díaz, Henry
Puentes-Garzón, David Esteban
Zarate-Caicedo, Diego Alejandro
author_facet Lamos-Díaz, Henry
Puentes-Garzón, David Esteban
Zarate-Caicedo, Diego Alejandro
author_sort Lamos-Díaz, Henry
collection OJS
description The identification of influencing factors in crop yield (kg·ha-1) provides essential information for decision-making processes related to the prediction and improvement of productivity, which gives farmers the opportunity to increase their income. The current study investigates the application of multiple machine learning algorithms for cocoa yield prediction and influencing factors identification. The Support Vector Machines (SVM) and Ensemble Learning Models (Random Forests, Gradient Boosting) are compared with Least Absolute Shrinkage and Selection Operator (LASSO) regression models. The considered predictors were climate conditions, cocoa variety, fertilization level and sun exposition in an experimental crop located in Rionegro, Santander. Results showed that Gradient Boosting is the best prediction alternative with Coefficient of determination (R2) = 68%, Mean Absolute Error (MAE) = 13.32, and Root Mean Square Error (RMSE) = 20.41. The crop yield variability is explained mainly by the radiation one month before harvest, the accumulated rainfall on the harvest month, and the temperature one month before harvest. Likewise, the crop yields are evaluated based on the kind of sun exposure, and it was found that radiation one month before harvest is the most influential factor in shade-grown plants. On the other hand, rainfall and soil moisture are determining variables in sun-grown plants, which is associated with the water requirements. These results suggest a differentiated management for crops depending on the kind of sun exposure to avoid compromising productivity, since there is no significant difference in the yield of both agricultural managements.
format Online
id oai:oai.revistas.uptc.edu.co:article-10853
institution Revista Facultad de Ingeniería
language eng
spa
publishDate 2020
publisher Universidad Pedagógica y Tecnológica de Colombia
record_format ojs
spelling oai:oai.revistas.uptc.edu.co:article-108532022-06-15T15:57:49Z Comparison Between Machine Learning Models for Yield Forecast in Cocoa Crops in Santander, Colombia Comparación de modelos de aprendizaje automático para la predicción de rendimientos agrícolas en cultivos de cacao en Santander, Colombia Lamos-Díaz, Henry Puentes-Garzón, David Esteban Zarate-Caicedo, Diego Alejandro agricultural-yield agroforestry-system cocoa machine-learning prediction productivity aprendizaje-automático cacao predicción productividad rendimientos-agrícolas sistemas-agroforestales The identification of influencing factors in crop yield (kg·ha-1) provides essential information for decision-making processes related to the prediction and improvement of productivity, which gives farmers the opportunity to increase their income. The current study investigates the application of multiple machine learning algorithms for cocoa yield prediction and influencing factors identification. The Support Vector Machines (SVM) and Ensemble Learning Models (Random Forests, Gradient Boosting) are compared with Least Absolute Shrinkage and Selection Operator (LASSO) regression models. The considered predictors were climate conditions, cocoa variety, fertilization level and sun exposition in an experimental crop located in Rionegro, Santander. Results showed that Gradient Boosting is the best prediction alternative with Coefficient of determination (R2) = 68%, Mean Absolute Error (MAE) = 13.32, and Root Mean Square Error (RMSE) = 20.41. The crop yield variability is explained mainly by the radiation one month before harvest, the accumulated rainfall on the harvest month, and the temperature one month before harvest. Likewise, the crop yields are evaluated based on the kind of sun exposure, and it was found that radiation one month before harvest is the most influential factor in shade-grown plants. On the other hand, rainfall and soil moisture are determining variables in sun-grown plants, which is associated with the water requirements. These results suggest a differentiated management for crops depending on the kind of sun exposure to avoid compromising productivity, since there is no significant difference in the yield of both agricultural managements. La identificación de los factores que influyen en el rendimiento (kg·ha-1) de un cultivo provee información esencial para la toma de decisiones orientadas al mejoramiento y predicción de la productividad, proporcionando posibilidades a los agricultores para mejorar sus ingresos económicos. En este estudio, se presenta la aplicación y comparación de diversos algoritmos de aprendizaje automático para la predicción del rendimiento agrícola en cultivos de cacao y la identificación de los factores que influyen sobre éste. Se comparan los algoritmos de máquinas de soporte vectorial (SVM), modelos ensamblados (Random Forest, Gradient Boosting) y el modelo de regresión Least Absolute Shrinkage and Selection Operator (LASSO). Los predictores considerados fueron: condiciones climáticas de la región, variedad de cacao, nivel de fertilización y exposición al sol para un cultivo experimental ubicado en Rionegro, Santander. Los resultados identifican a Gradient Boosting como la mejor alternativa de pronóstico con un coeficiente de determinación (R2) = 68 %, Error Absoluto Medio (MAE) = 13.32 y Raíz Cuadrada del Error Medio (RMSE) = 20.41. La variabilidad del rendimiento del cultivo es explicada principalmente por la radiación y la temperatura un mes previo a la cosecha, además de las lluvias acumuladas el mes de la cosecha. De igual manera, los rendimientos de los cultivos son evaluados con base en el tipo de exposición al sol, encontrando que la radiación un mes previo a la cosecha es el factor más influyente para los cultivos bajo sombra. Por otro lado, la lluvia y la humedad son las variables determinantes en las plantas con exposición plena a sol, lo que está asociado a los requerimientos hídricos. Estos resultados sugieren un manejo diferenciado de los cultivos dependiendo del tipo de exposición, sin tener que comprometer la productividad, dado que no se evidencia diferencia significativa en los rendimientos de ambos manejos agrícolas. Universidad Pedagógica y Tecnológica de Colombia 2020-05-15 info:eu-repo/semantics/article info:eu-repo/semantics/publishedVersion application/pdf application/pdf application/xml https://revistas.uptc.edu.co/index.php/ingenieria/article/view/10853 10.19053/01211129.v29.n54.2020.10853 Revista Facultad de Ingeniería; Vol. 29 No. 54 (2020): Continuos Publication; e10853 Revista Facultad de Ingeniería; Vol. 29 Núm. 54 (2020): Publicación Continua; e10853 2357-5328 0121-1129 eng spa https://revistas.uptc.edu.co/index.php/ingenieria/article/view/10853/9281 https://revistas.uptc.edu.co/index.php/ingenieria/article/view/10853/9282 https://revistas.uptc.edu.co/index.php/ingenieria/article/view/10853/9507 Copyright (c) 2020 Henry Lamos-Díaz; David-Esteban Puentes-Garzón; Diego-Alejandro Zarate-Caicedo
spellingShingle agricultural-yield
agroforestry-system
cocoa
machine-learning
prediction
productivity
aprendizaje-automático
cacao
predicción
productividad
rendimientos-agrícolas
sistemas-agroforestales
Lamos-Díaz, Henry
Puentes-Garzón, David Esteban
Zarate-Caicedo, Diego Alejandro
Comparison Between Machine Learning Models for Yield Forecast in Cocoa Crops in Santander, Colombia
title Comparison Between Machine Learning Models for Yield Forecast in Cocoa Crops in Santander, Colombia
title_alt Comparación de modelos de aprendizaje automático para la predicción de rendimientos agrícolas en cultivos de cacao en Santander, Colombia
title_full Comparison Between Machine Learning Models for Yield Forecast in Cocoa Crops in Santander, Colombia
title_fullStr Comparison Between Machine Learning Models for Yield Forecast in Cocoa Crops in Santander, Colombia
title_full_unstemmed Comparison Between Machine Learning Models for Yield Forecast in Cocoa Crops in Santander, Colombia
title_short Comparison Between Machine Learning Models for Yield Forecast in Cocoa Crops in Santander, Colombia
title_sort comparison between machine learning models for yield forecast in cocoa crops in santander colombia
topic agricultural-yield
agroforestry-system
cocoa
machine-learning
prediction
productivity
aprendizaje-automático
cacao
predicción
productividad
rendimientos-agrícolas
sistemas-agroforestales
topic_facet agricultural-yield
agroforestry-system
cocoa
machine-learning
prediction
productivity
aprendizaje-automático
cacao
predicción
productividad
rendimientos-agrícolas
sistemas-agroforestales
url https://revistas.uptc.edu.co/index.php/ingenieria/article/view/10853
work_keys_str_mv AT lamosdiazhenry comparisonbetweenmachinelearningmodelsforyieldforecastincocoacropsinsantandercolombia
AT puentesgarzondavidesteban comparisonbetweenmachinelearningmodelsforyieldforecastincocoacropsinsantandercolombia
AT zaratecaicedodiegoalejandro comparisonbetweenmachinelearningmodelsforyieldforecastincocoacropsinsantandercolombia
AT lamosdiazhenry comparaciondemodelosdeaprendizajeautomaticoparalaprediccionderendimientosagricolasencultivosdecacaoensantandercolombia
AT puentesgarzondavidesteban comparaciondemodelosdeaprendizajeautomaticoparalaprediccionderendimientosagricolasencultivosdecacaoensantandercolombia
AT zaratecaicedodiegoalejandro comparaciondemodelosdeaprendizajeautomaticoparalaprediccionderendimientosagricolasencultivosdecacaoensantandercolombia