• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.11.1983.tde-20220207-193849
Document
Author
Full name
Elisabeth Borges Goncalves
Institute/School/College
Knowledge Area
Date of Defense
Published
Piracicaba, 1983
Supervisor
Title in Portuguese
Comparação de modelos para previsão de series temporais -serie de precipitações pluviais mensais no município de Pindorama -SP
Keywords in Portuguese
CHUVA
MODELOS EM SÉRIES TEMPORAIS
PREVISÃO
Abstract in Portuguese
O trabalho focaliza análise de series temporais no domínio do tempo, a partir de conceitos básicos em modelagem e previsão, apresentando-se um sistema preditivo. As classes de modelos Naive, Suavizados, Autoregressivos Box e Jenkins e métodos de decomposição são desenvolvidos e discutidos em seus aspectos prático e teórico. No intuito de estudar estes modelos em series climatológicas e pesquisar sua acuracidade em casos excessivamente aleatórios, é analisada a série de precipitações pluviais mensais no Município de Pindorama-SP, a partir de janeiro de 1950 a abril de 1981, discutindo-se a performance dos modelos. Como instrumentos para tomadas de decisão na escolha entre modelos autoregressivos e suavizados são propostas e discutidas as medidas ARSE e SNR, utilizando-se series econômicas de domínio público (oito) e processos ARMA simulados no intuito de verificar a adequabilidade destas medidas. De modo geral, os melhores modelos foram o de suavização sazonal aditiva de Holt e Winters e ARMA sazonal, com relação ao EQM de previsão, porem enquanto o suavizado não apresentou f.a.c. residual aleatória, o ARMA ajustado revelou-se pouco prático. As medidas ARSE e SNR obtiveram um desempenho razoável, mas ainda devem ser melhor desenvolvidas, o que envolverá discussões· sobre o sentido de aleatoriedade em series temporais.
Title in English
Comparation of forecasting m odels in time series monthly pluvial precipitations in Pindorama -SP
Abstract in English
This study comprises time series analysis on time domain, beginni'ng wtth concepts in models construction and forecastingi being introduced a predictive system. Smoothing methods, Naive models, Autoregressive, ARIMA and Decomposition technics are developed and discussed in their pratical and theoretical aspects. To observe this models on climatological series and research their accuracy in randon cases, a pluvial time serie is analysed, the serie is composed by 364 monthly observations from january, 1950 to april, 1981 in Pindorama-SP, looking upon the models performance. As instruments to make decision to pick out autoregressive , or smoothing models are proposed the measures ARSE and SNR, using public. economical series (eight series) and simulated ARMA processes to verify their properties. ln a general way, the best models were the Holt-Winters' adi tive seasonal smoothing and the seasonal ARMA model, relative to M.S.E., but the smoothing method had not random residual autocorrelation function and the ARMA wasnt practical. Measures ARSE and SNR had a reasonable accomplishment, but they must be something more developed, what will involve considerations about the sense of random components in time series.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2022-02-07
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2022. All rights reserved.