The Rural Environmental Registry (CAR) consists of a mandatory public electronic registry for all rural properties in the Brazilian territory, integrates environmental information of the properties, assists the monitoring of them, and the fight against deforestation. However, a large number of registrations are carried out erroneously generating inconsistent data, leading these to be canceled and/or to be requested to correct the registration. Carrying out these checks manually is very expensive, since a specialized workforce is required and Brazil has an immense amount of rural properties. In this context, this work aims to provide an intelligent machine learning-based system that allows to verify and classify CAR data into approved or canceled data quickly and effectively. For this purpose, three learning models were trained using real data from registers. In addition to the classification, the SMOTE tool was used to treat the imbalance between classes. Results were generated using measures of performance of classifiers and comparative studies between the methods were also performed. The results showed potential use of the method in future automated predictions, reaching performance indices above 0.90 (90%). Resumo: O Cadastro Ambiental Rural (CAR) consiste em um registro público eletrônico obrigatório para todos os imóveis rurais do território brasileiro, integra informações ambientais das propriedades, auxilia o monitoramento das mesmas e no combate ao desmatamento. Entretanto, um grande número de cadastros é realizado de maneira errônea gerando dados inconsistentes, levando estes a serem cancelados e/ou a serem pedidas retificações para o devido preenchimento do cadastro. Realizar essas verificações de forma manual é deveras oneroso, uma vez que é requerida uma mão de obra especializada e o Brasil possui uma imensa quantidade de imóveis rurais. Neste contexto, este trabalho tem como objetivo fornecer um sistema inteligente baseado em aprendizagem de máquina que permita verificar e classificar os dados do CAR em aprovados ou cancelados de maneira rápida e eficaz. Para isto, três modelos de aprendizagem foram treinados utilizando dados reais de cadastros. Além da classificação, foi utilizada a ferramenta SMOTE para tratamento do desbalanceamento entre as classes. Foram gerados resultados utilizando medidas de desempenho de classificadores e realizados, também, estudos comparativos entre os métodos. Os resultados apresentados mostraram potencial uso do método em futuras predições automatizadas, atingindo índices de desempenho acima de 0.90 (90%).
The Rural Environmental Registry (CAR) consists of a mandatory public electronic registry for all rural properties in the Brazilian territory, integrates environmental information of the properties, assists the monitoring of them and the fight against deforestation. However, a large number of registrations are carried out erroneously generating inconsistent data, leading these to be canceled and/or to be requested to correct the registration. Performing automatic verification of these records is important to improve the processing of records. This paper proposes an automatic classification method to approve or cancel the CAR registers with interpretation of the classifications performed. For this, four machine learning-based classifiers were tested and the results were evaluated. The model with the best performance was used to interpret the classification using the Local Interpretable Model-agnostic Explanations (LIME) algorithm. The results showed the potential of the method in future real applications.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.