bsts_modelspec: BSTS Model Secification

In tsmodels/tsforeign: Interface To Other Packages

BSTS Model Secification

Description

Specifies a BSTS model prior to estimation.

Usage

bsts_modelspec(
  y,
  xreg = NULL,
  frequency = NULL,
  differences = 0,
  level = TRUE,
  slope = TRUE,
  damped = FALSE,
  seasonal = FALSE,
  seasonal_frequency = 4,
  ar = FALSE,
  ar_max = 1,
  cycle = FALSE,
  cycle_frequency = NULL,
  cycle_names = NULL,
  seasonal_type = "regular",
  seasonal_harmonics = NULL,
  transformation = "box-cox",
  lambda = NULL,
  lower = 0,
  upper = 1,
  distribution = "gaussian",
  ...
)

Arguments

y	an xts vector.
xreg	an xts matrix of external regressors.
frequency	frequency of y (if using a seasonal model).
differences	number of differences to apply to the outcome variable y (max of 2).
level	whether to include a level component (Local Level Model).
slope	whether to include a slope component (Local Linear Model).
damped	whether to include a damped trend (damped Local Linear Model).
seasonal	whether to include a seasonal component.
seasonal_frequency	vector of seasonal frequencies.
ar	whether to include a sparse AR component.
ar_max	number of lags for the AR component.
cycle	whether to include a cyclical component.
cycle_frequency	number of periods in a cycle. This can be a vector in which case multiple cycles are included.
cycle_names	optional vector of cycle names.
seasonal_type	type of seasonality (regular or trigonometric).
seasonal_harmonics	number of harmonics to include in the seasonal component when seasonal_type is trigonometric.
transformation	a valid transformation for y from the “tstransform” function in the “tsaux” package (currently box-cox or logit are available).
lambda	the Box Cox lambda. If NA will estimate this using the method of Guerrero.
lower	lower bound for the transformation.
upper	upper bound for the transformation.
distribution	valid choices are currently only “gaussian”.
...	not currently used.

Value

An object of class “bsts.spec”.

Note

This is a wrapper to part of the functionality of the bsts package. Once an object is estimated, all other methods are implemented locally (including prediction).

tsmodels/tsforeign documentation built on June 22, 2022, 2:09 p.m.