PROCESSOS k-FACTOR GARMA: PROPRIEDADES, SIMULAÇÕES E APLICAÇÃO
PROCESSOS k-FACTOR GARMA: PROPRIEDADES, SIMULAÇÕES E APLICAÇÃO
DOI: 10.22533/at.ed.7782317104
Palavras-chave: Série temporal, processos k-Factor GARMA(p,λ,u,q), longa dependência, estimação, aplicação
Keywords: Time series, k-Factor GARMA(p,λ,u,q) processes, long-range dependence, estimation, application
Abstract: The main objective of this work is to study k-Factor GARMA (p,λ,u,q) processes. These processes exhibit the characteristic of long dependence and are more general than ARFIMA (p,d,q) and SARFIMA (p,d,q)× P,D,Q s processes. We present significant results related to stationarity, invertibility conditions, infinite autoregressive and moving average representations, their spectral density function, and their behavior near Gegenbauer frequencies. We also provide an estimator for Gegenbauer frequencies and two parametric estimators for process parameters, namely Whittle and CSS. Through Monte Carlo simulations, we investigate the performance of the parametric estimators, confirming that these estimators are unbiased and consistent. Finally, we conduct an analysis of the time series of the Southern Oscillation Index (SOI), an indicator of the Southern Oscillation state. This analysis reveals negative values of the SOI index, which are associated with the El Niño phenomenon.
- Cleber Bisognin
- Caroline Lopes Gonçalves
- Keler Eliana Severo Correa