Hybrid Models Applied to Create a Classification Index of Fire Risk Levels in Brazil

• Autor(a) principal: Galvao Junior, Pedro Antonio
• Data de Publicação: 2022
• Outros Autores: Roveda, Sandra Regina Monteiro Masalskiene, Vieira, Henrique Ewbank de Miranda
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Revista Brasileira de Ciências Ambientais (Online)
• Texto Completo: https://www.rbciamb.com.br/Publicacoes_RBCIAMB/article/view/1286
• topic: fuzzy modeling; forecast model; machine learning; neurofuzzy model; artificial neural networks.; modelagem fuzzy; modelo de previsão; machine learning; modelo neuro-fuzzy; redes neurais artificias.

Fire has always exerted a great attraction on humans. Fires generally provide social and environmental impacts at the places where they occur. Several Brazilian localities, especially in the driest months of the year, are more susceptible to this phenomenon. In this paper, an index able of classifying levels of fire risk in areas geographically located in Brazil. This paper presents an index capable of classifying fire risk levels elaborated from neuro-fuzzy systems. Data from the municipality of Sorocaba were used to test the proposed models. The results obtained by this index are promising, reaching values of mean absolute error below 3% when applied in the prediction of the risk of fire for the maximum period of up to 3 days. The proposed index can be used as a tool to support and assist various research agencies or institutes that need to identify the possibility of burning, corroborating the measures to reduce atmospheric emitters and meeting Goal 15 of Agenda 30 as defined by the UN in 2015, which aims to stimulate conservation actions and the recovery and sustainable use of ecosystems.