• Entrar
    Ver item 
    •   Página inicial
    • BIBLIOTECA DIGITAL: Teses & Dissertações
    • 40001016075P3 Programa de Pós-Graduação em Engenharia Ambiental
    • Dissertações
    • Ver item
    •   Página inicial
    • BIBLIOTECA DIGITAL: Teses & Dissertações
    • 40001016075P3 Programa de Pós-Graduação em Engenharia Ambiental
    • Dissertações
    • Ver item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Emprego de redes neurais artificiais na previsão climática de temperatura e precipitação no Estado do Paraná

    Thumbnail
    Visualizar/Abrir
    R - D - CARLA MILLEO.pdf (3.950Mb)
    Data
    2020
    Autor
    Milléo, Carla, 1993-
    Metadata
    Mostrar registro completo
    Resumo
    Resumo: O objetivo deste estudo e avaliar o desempenho das Redes Neurais Artificiais (RNA) de contra-propagacao (CPN) e de funcao de base radial (RBF) na previsao climatica de temperatura e precipitacao no estado do Parana. Uma etapa importante no desenvolvimento de um modelo de RNA e a selecao de variaveis de entrada. Para isso, buscou-se na literatura os principais fenomenos atmosfericos que produzem mudancas nos quadros medios de temperatura e precipitacao no Parana. Apos selecionar os possiveis preditores das variacoes climaticas, foi utilizado o metodo de regressao stepwise, passo forward, para identificar o melhor conjunto de variaveis de entrada das redes neurais. A rede neural CPN, com 150 neuronios na camada oculta, e RBF, com 50 neuronios na camada oculta, foram desenvolvidas para cada variavel a ser prevista. Os resultados mostraram que o desempenho da rede do tipo RBF foi superior ao da rede de CPN em todos os experimentos de previsoes. Os melhores desempenhos foram obtidos para as previsoes de variaveis filtradas por uma media movel de tres meses, em virtude de dados que possuem muito ruido sao mais dificeis de serem simulados por RNA. As tecnicas de RNAs mostraram melhor desempenho nas estimativas de temperatura e precipitacao do Parana que a tecnica de persistencia. Palavras-chave: Previsao de temperatura e precipitacao. Redes Neurais Artificiais. Selecao de variaveis de entrada. Redes Neurais de Contra-Propagacao. Redes Neurais RBF.
     
    Abstract: The objective of this study is to evaluate the performance of counter-propagation (CPN) and radial base function (RBF) Artificial Neural Networks (ANN) in the climatic forecast of temperature and precipitation in the state of Parana. An important step in the development of an ANN model is the selection of input variables. For that, the main atmospheric phenomena that produced changes in the average temperature and precipitation conditions in Parana were searched in the literature. After selecting the possible predictors of climatic variations, the stepwise regression method, step forward, was used to identify the best set of input variables for neural networks. The CPN neural network, with 150 neurons in the hidden layer, and RBF, with 50 neurons in the hidden layer, were developed for each variable to be predicted. The results showed that the performance of the RBF network was superior to that of the CPN network in all forecasting experiments. The best performances were obtained for the predictions of variables filtered by a three-month moving average, due to data that have a lot of noise are more difficult to be simulated by RNA. The ANN techniques performed better than simple persistence in the forecast of temperature and precipitation for state of Parana. Keywords: Temperature and precipitation forecasting. Artificial Neural Network. Selection of input variables. Counterpropagation Neural Networks. RBF Neural Networks
     
    URI
    https://hdl.handle.net/1884/69952
    Collections
    • Dissertações [80]

    DSpace software copyright © 2002-2022  LYRASIS
    Entre em contato | Deixe sua opinião
    Theme by 
    Atmire NV
     

     

    Navegar

    Todo o repositórioComunidades e ColeçõesPor data do documentoAutoresTítulosAssuntosTipoEsta coleçãoPor data do documentoAutoresTítulosAssuntosTipo

    Minha conta

    EntrarCadastro

    Estatística

    Ver as estatísticas de uso

    DSpace software copyright © 2002-2022  LYRASIS
    Entre em contato | Deixe sua opinião
    Theme by 
    Atmire NV