Incorporating soil information with machine learning for crop recommendation to improve agricultural output

Afzal, Hadeeqa and Amjad, Madiha and Raza, Ali and Munir, Kashif and Gracia Villar, Santos and Dzul López, Luis Alonso and Ashraf, Imran UNSPECIFIED, UNSPECIFIED, UNSPECIFIED, UNSPECIFIED, santos.gracia@uneatlantico.es, luis.dzul@uneatlantico.es, UNSPECIFIED (2025) Incorporating soil information with machine learning for crop recommendation to improve agricultural output. Scientific Reports, 15 (1). ISSN 2045-2322

Text s41598-025-88676-z.pdf Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (3MB)

Abstract

The agriculture field is the basis of a country’s change and financial system. Crops are the main source of revenue for the people. One of the farmer’s most challenging problems is choosing the right crops for their land. This critical decision has a direct impact on productivity and profit. Wrong crop selection not only reduces yields but also causes food shortages, creating more problems for farmers. The best crop depends on many parameters such as illustration humidity, N, K, P, pH, rainfall, and temperature of the soil. Getting advice from experts is not an easy task. This requires intelligent models in crop recommendations that use machine-learning models to suggest suitable crops for soil and other environmental conditions. Temperature, humidity, and pH are important data for growing crops in agriculture. In this study, we gather and preprocess relevant data. To recommend the most suitable crop, we propose a novel ensemble learning approach called RFXG based on random forest (RF) and extreme gradient boosting (XGB) to suggest the best crop out of the twenty-two major crops. To measure the capability of the proposed approach, various machine learning models are utilized including extra tree classifier, multilayer perceptron, RF, decision trees, logistic regression, and XGB classifiers. To get the best performance, optimization of hyperparameter, and K-fold cross-validation procedures are performed. Experimental outcomes show that the proposed RFXG technique achieves a recommendation accuracy is 98%. Specifically, the proposed solution provides immediate recommendations to help farmers make timely decisions.

Item Type:	Article
Uncontrolled Keywords:	Agricultural economics, Smart agriculture, Soil parameters, Crop recommendation, Smart farming, Machine learning, Ensemble model
Subjects:	Subjects > Engineering
Divisions:	Europe University of Atlantic > Research > Scientific Production Ibero-american International University > Research > Scientific Production Universidad Internacional do Cuanza > Teaching > Teaching Materials University of La Romana > Research > Scientific Production
Depositing User:	Sr Bibliotecario
Date Deposited:	17 Mar 2025 09:15
Last Modified:	17 Mar 2025 09:15
URI:	http://repositorio.funiber.org/id/eprint/17272

Actions (login required)

View Item