season: Seasonal Adjustment

In TSSS: Time Series Analysis with State Space Model

Seasonal Adjustment

Description

Seasonal adjustment by state space modeling.

Usage

season(y, trend.order = 1, seasonal.order = 1, ar.order = 0, trade = FALSE,
       period = NULL, tau2.ini = NULL, filter = c(1, length(y)),
       predict = length(y), arcoef.ini = NULL, log = FALSE, log.base = "e",
       minmax = c(-1.0e+30, 1.0e+30), plot = TRUE, ...)

Arguments

y	a univariate time series with or without the tsp attribute.
trend.order	trend order (0, 1, 2 or 3).
seasonal.order	seasonal order (0, 1 or 2).
ar.order	AR order (0, 1, 2, 3, 4 or 5).
trade	logical; if TRUE, the model including trading day effect component is considered.
period	If the tsp attribute of y is NULL, valid number of seasons in one period in the case that seasonal.order > 0 and/or trade = TRUE. 4 : quarterly data 12 : monthly data 5 : daily data (5 days a week) 7 : daily data (7 days a week) 24 : hourly data
tau2.ini	initial estimate of variance of the system noise \tau^2 less than 1.
filter	a numerical vector of the form c(x1,x2) which gives start and end position of filtering.
predict	the end position of prediction (\geq x2).
arcoef.ini	initial estimate of AR coefficients (for ar.order > 0).
log	logical. If TRUE, the data y is log-transformed.
log.base	the letter "e" (default) or "10" specifying the base of logarithmic transformation. Valid only if log = TRUE.
minmax	lower and upper limits of observations.
plot	logical. If TRUE (default), trend, seasonal, AR and noise components are plotted.
...	graphical arguments passed to plot.season.

Value

An object of class "season", which is a list with the following components:

tau2	variance of the system noise.
sigma2	variance of the observational noise.
llkhood	log-likelihood of the model.
aic	AIC of the model.
trend	trend component (for trend.order > 0).
seasonal	seasonal component (for seasonal.order > 0).
arcoef	AR coefficients (for ar.order > 0).
ar	AR component (for ar.order > 0).
day.effect	trading day effect (for trade = TRUE).
noise	noise component.
cov	covariance matrix of smoother.

Note

For time series with the tsp attribute, set frequency to period.

References

Kitagawa, G. (2020) Introduction to Time Series Modeling with Applications in R. Chapman & Hall/CRC.

Examples

# BLSALLFOOD data
data(BLSALLFOOD)
season(BLSALLFOOD, trend.order = 2, seasonal.order = 1, ar.order = 2)

season(BLSALLFOOD, trend.order = 2, seasonal.order = 1, ar.order = 2,
       filter = c(1, 132))

# Wholesale hardware data
data(WHARD)
season(WHARD, trend.order = 2, seasonal.order = 1, ar.order = 0, trade = TRUE,
       log = TRUE)

season(WHARD, trend.order = 2, seasonal.order = 1, ar.order = 0, trade = TRUE,
       filter = c(1, 132), log = TRUE)

TSSS documentation built on Sept. 29, 2023, 9:07 a.m.