determine_beta: Helpfunction in estimate_beta() for estimating beta_est.

Description

Helpfunction in estimate_beta() for estimating beta_est.

Usage

determine_beta(
  string,
  X_special,
  Y_special,
  robust,
  NN,
  TT,
  S,
  method_estimate_beta,
  initialisation = FALSE,
  indices = NA,
  vars_est,
  sigma2,
  nosetting_local = FALSE,
  kappa_candidates = c(2^(-0:-20), 0)
)

Value

The function returns a numeric vector (for the default setting: string == "heterogeneous") or a matrix with the estimated beta (if string == "homogeneous").

Arguments

string: can have values: "homogeneous" (when one beta_est is estimated for all individuals together) or "heterogeneous" (when beta_est is estimated either groupwise or elementwise)
X_special: preprocessed X (2-dimensional matrix with 'var_est' observable variables)
Y_special: preprocessed Y
robust: robust or classical estimation
NN: number of time series
TT: length of time series
S: estimated number of groups
method_estimate_beta: defines how beta is estimated. Default case is an estimated beta for each individual. Default value is "individual." Possible values are "homogeneous", "group" or "individual".
initialisation: indicator of being in the initialisation phase
indices: individuals for which beta_est is being estimated
vars_est: number of available observed variables for which a coefficient will be estimated. As default it is equal to the number of available observed variables.
sigma2: sum of squared error terms, scaled by NT
nosetting_local: option to remove the recommended setting in lmrob(). It is much faster. Defaults to FALSE.
kappa_candidates: Defines the size of the SCAD-penalty used in the classical algorithm. This vector should contain more than 1 element.