Estimate of Acacia mangium volume using techniques of artificial neural networks and support vector machines

• Autor(a) principal: Cordeiro, Márcio Assis
• Data de Publicação: 2015
• Outros Autores: Pereira, Nayara Natacha de Jesus, Binoti, Daniel Henrique Breda, Binoti, Mayra Luiza Marques da Silva, Leite, Hélio Garcia
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Pesquisa Florestal Brasileira (Online)
• Texto Completo: https://pfb.cnpf.embrapa.br/pfb/index.php/pfb/article/view/596
• topic: Smalian; Modelagem; Estimativa volumétrica; Smalian; Modeling; Volumetric estimates

The present study aimed to show the results of Acacia mangium volumetric estimates obtained through the Schumacher and Hall model compared to the methods of artificial neural networks and support vector machines. To enable this comparative analysis, we used data from 31 trees of Acacia mangium aged 14–17, from a stand located in the northern region of the state of Amapá. Diameter and bark thickness of the trees were measured into relative heights along the stem into 14 sections (0.05%, 1%, 5%, 10%, 15%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, and 95%), with measurement. Total volume with bark was obtained by applying the Smalian formula. In general, the methods that differ from traditional methods showed statistically superior results.