dc.contributor | Rezende, Leiliane Pereira de | |
dc.contributor | Miranda, Glauco Vieira | |
dc.contributor | Rezende, Leiliane Pereira de | |
dc.contributor | Miranda, Glauco Vieira | |
dc.contributor | Naves, Thiago França | |
dc.contributor | Sepulveda, Gloria Patricia Lopez | |
dc.creator | Souza, Paulo Vitor Duarte de | |
dc.date.accessioned | 2021-11-18T15:53:07Z | |
dc.date.accessioned | 2022-12-06T14:39:15Z | |
dc.date.available | 2021-11-18T15:53:07Z | |
dc.date.available | 2022-12-06T14:39:15Z | |
dc.date.created | 2021-11-18T15:53:07Z | |
dc.date.issued | 2021-08-23 | |
dc.identifier | SOUZA, Paulo Vitor Duarte de. Rede neural artificial para predição da produtividade da cultura do milho. 2021. Trabalho de Conclusão de Curso (Bacharelado em Ciência da Computação) - Universidade Tecnológica Federal do Paraná, Santa Helena, 2021. | |
dc.identifier | http://repositorio.utfpr.edu.br/jspui/handle/1/26421 | |
dc.identifier.uri | https://repositorioslatinoamericanos.uchile.cl/handle/2250/5253365 | |
dc.description.abstract | The forecasting corn yield has many advantages in global food production and small farmers. The forecasting can manage better the crop and optimize the profit. The objective of the work is to build models of Multilayer Perceptrons for the forecasting of corn yield in the Vale do Paranapanema, São Paulo, with the crop growing fe atures, weather conditions and water balance. Two year (2018 and 2019) crop of different locations were considered. Missing data were imputed through the iterative imputation. The model hyper parametrization were obtained through of the Grid Search and k-fold cross-validation. The models were divided into datasets with and without data imputation. The interpretation of the model was made through SHAP method. The models obtained satisfactory results, with the better model had RMSE de 70,651 kg ·ha−1 for the imputed dataset. In the data without imputation obtained 190,851 kg ·ha−1. In all models, the weather conditions were that had importance in predictions of yield. Thus, the build models had satisfactory performance and they took the nonlinear interactions among the crop genotype and the environment. | |
dc.publisher | Universidade Tecnológica Federal do Paraná | |
dc.publisher | Santa Helena | |
dc.publisher | Brasil | |
dc.publisher | Ciência da Computação | |
dc.publisher | UTFPR | |
dc.rights | openAccess | |
dc.subject | Redes neurais (Computação) | |
dc.subject | Milho | |
dc.subject | Produtividade agrícola | |
dc.subject | Neural networks (Computer science) | |
dc.subject | Corn | |
dc.subject | Agricultural productivity | |
dc.title | Rede neural artificial para predição da produtividade da cultura do milho | |
dc.type | bachelorThesis | |