Registro completo |
Provedor de dados: |
PFB - Pesquisa Florestal Brasileira
|
País: |
Brazil
|
Título: |
Descrição do perfil do tronco de árvores em plantios de diferentes espécies por meio de redes neurais artificiais
Stem profile description in plantations for different species using artificial neural network
|
Autores: |
Campos, Bráulio Pizziôlo Furtado
Silva, Gilson Fernandes da
Binoti, Daniel Henrique Breda
Mendonça, Adriano Ribeiro de
Leite, Helio Garcia
|
Data: |
2017-06-30
|
Ano: |
2017
|
Palavras-chave: |
Inventário florestal
Modelos de Crescimento e Produção
Estatística Inventário Florestal
Manejo Florestal
Inteligência artificial Forest inventory
Forest management
Artificial intelligence
|
Resumo: |
The objective of this study was to analyze the ability of an artificial neural network (ANN) to describe the stem profile of trees of different genera and species in different growing conditions. For comparative purposes, equations were fit, using regression analysis to describe the stem profile. For neural network as well as for the regression equations, evaluation of accuracy was based on correlation coefficient between observed and estimated diameters along the stem, square root of the mean square percentage error (RMSE) and graphical analysis. Artificial intelligence methods, especially ANN, can be effective in describing trees bole profile of different species in different growth conditions using only one ANN with similar efficiency as regression models traditionally employed by forestry companies.
The objective of this study was to analyze the ability of an artificial neural network (ANN) to describe the stem profile of trees of different genera and species in different growing conditions. For comparative purposes, equations were fit, using regression analysis to describe the stem profile. For neural network as well as for the regression equations, evaluation of accuracy was based on correlation coefficient between observed and estimated diameters along the stem, square root of the mean square percentage error (RMSE) and graphical analysis. Artificial intelligence methods, especially ANN, can be effective in describing trees bole profile of different species in different growth conditions using only one ANN with similar efficiency as regression models traditionally employed by forestry companies.
|
Tipo: |
Info:eu-repo/semantics/article
|
Idioma: |
Português
|
Identificador: |
http://pfb.cnpf.embrapa.br/pfb/index.php/pfb/article/view/1181
10.4336/2017.pfb.37.90.1181
|
Editor: |
Embrapa Florestas
|
Relação: |
http://pfb.cnpf.embrapa.br/pfb/index.php/pfb/article/view/1181/564
http://pfb.cnpf.embrapa.br/pfb/index.php/pfb/article/downloadSuppFile/1181/901
|
Formato: |
application/pdf
|
Fonte: |
Pesquisa Florestal Brasileira; v. 37, n. 90 (2017): abr./jun.; 99-107
Brazilian Journal of Forestry Research; v. 37, n. 90 (2017): abr./jun.; 99-107
1983-2605
1809-3647
|
Direitos: |
http://creativecommons.org/licenses/by-nc-nd/4.0
|