trend-decomp-method: Obtain Individual Components of a Decomposed Time Series

In deseats: Data-Driven Locally Weighted Regression for Trend and Seasonality in TS

Obtain Individual Components of a Decomposed Time Series

Description

The provided methods work for decomposition objects created within this package. They allow the user to retrieve individual components among the estimated ones.

Usage

## S4 method for signature 'decomp'
trend(object, ...)

## S4 method for signature 'decomp'
season(object, ...)

## S4 method for signature 'decomp'
fitted(object, ...)

## S4 method for signature 'decomp'
residuals(object, ...)

## S4 method for signature 'decomp'
deseasonalize(object, ...)

## S4 method for signature 'decomp'
detrend(object, ...)

Arguments

object	the estimation / decomposition object, for example of class "deseats".
...	without further use; implemented for compatibility only.

Details

Apply these functions directly to an estimation object, i.e. the result of a decomposition of a seasonal time series, in order to retrieve individual estimated components.

Value

These methods return time series objects of class "ts" that represent the corresponding estimated component in the time series originally used in the estimation process.

Examples

Xt <- log(EXPENDITURES)
smoothing_options <- set_options(order_poly = 3)
est <- deseats(Xt, smoothing_options = smoothing_options)
trend_e <- trend(est)           # Trend estimates
season_e <- season(est)         # Seasonality estimates
trend_season_e <- fitted(est)   # Trend + seasonality estimates
resid_e <- residuals(est)       # Residuals (observ. - trend - seasonality)
ts_adj <- deseasonalize(est)    # Seasonally adjusted series
ts_notrend <- detrend(est)      # Detrended series

deseats documentation built on June 23, 2025, 5:09 p.m.