Modelagem fuzzy de dados climáticos estimados por modelo matemático - CCATT-BRAMS

• Ano de defesa: 2015
• Autor(a) principal: Vieira, Luciana Cristina Pompeo Ferreira da Silva [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Ar - Poluição; Lógica difusa; Aparelho respiratorio - Doenças; Modelos matematicos; Fuzzy logic
• Link de acesso: http://hdl.handle.net/11449/139420; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/18-05-2016/000863376.pdf

Currently, studies on air quality has been of utmost importance because the pollutants in the air cause direct effects on the respiratory system causing social costs. Given the importance of providing hospitalization for respiratory diseases to the city manager can functionally prepare the health service for possible admissions, this work aimed to develop and validate a fuzzy linguistic model for prediction of hospitalization for respiratory diseases. A fuzzy model was built to predict hospitalizations for pneumonia, bronchitis, bronchiolitis and asthma (J12 to J18 and J45 of ICD 10) after exposure to PM2.5 and CO in Volta Redonda residents, RJ, in 2012. They were built three models: two with two Entries (PM2.5 / TEMP and CO / TEMP) and one with three inputs (PM 2.5 / CO / TEMP). For the model with two entrances there were three membership functions for each, for PM2.5 concentrations or CO and temperature, as well as, an outlet with three membership functions for admissions, which were obtained from DATASUS. For the model with three inputs, membership functions were maintained and the output changed to four membership functions. The models showed good accuracy. For the model with PM2.5 the result was between 90% and 76.5% for lags 1, 2 and 3; for model with CO the result was between 94.7% and 80.1%, for lags 1, 2 and 3; and the model with both CO and PM2.5 accuracy was between 91.4% and 80.1% (PM2.5) and 92.9% and 82.1% (CO) for lags 1, 2 and 3, allowing its application for health managers