stsm_init_pars: Get initial parameter estimates for estimation

Description

Get initial parameter estimates for estimation

Usage

stsm_init_pars(
  y,
  freq,
  trend,
  cycle,
  decomp = "",
  seasons = NULL,
  prior = NULL,
  sig_level = 0.01,
  arma = c(p = NA, q = NA),
  exo = NULL,
  state_eqns = NULL,
  interpolate = NA,
  interpolate_method = NA
)

Value

named vector containing the initial parameter estimates for estimation

Arguments

y: an object created from stsm_detect_frequency
freq: Frequency of the data
trend: Trend specification ("random-walk", "random-walk-drift", "double-random-walk", "random-walk2").
cycle: The period for the longer-term cycle
decomp: Decomposition model ("tend-cycle-seasonal", "trend-seasonal", "trend-cycle", "trend-noise")
seasons: The seasonal lengths to split the seasonality into
prior: A data table created by stsm_prior
sig_level: Significance level for statistical tests
arma: Named vector with values for p and q corresponding to the ARMA(p,q) specification if
exo: Matrix of exogenous variables. Can be used to specify regression effects or other seasonal effects like holidays, etc.
state_eqns: Character vector of equations to apply exo_state to the unobserved components. If left as the default, then all variables in exo_state will be applied to all the unobserved components. The equations should look like: "trend ~ var - 1", "drift ~ var - 1", "cycle ~ var - 1", "seasonal ~ var - 1". If only some equations are specified, it will be assumed that the exogenous data will be applied to only those specified equations.
interpolate: Character string giving frequency to interpolate to: i.e. "quarterly", "monthly", "weekly", "daily" cycle is set to 'arma'. If NA, then will auto-select the order.
interpolate_method: Character string giving the interpolation method: