prior_a: Determine Scale Factor Paramenter "a" for Laplace...

In nelson16silva/wavdrcast: Wavelet-Based Direct Multistep Forecast

Description

The function prior_a determines the parameter a of the laplace distribution that is used in empirical Bayes thresholding approach. Estimation is based on the avarage of the forecast error.

Usage

prior_a(..., wt = "modwt", interval = c(0.1, 5), tol = 0.01)

Arguments

...	Parameters for function error. Argument wav = TRUE is automatically adjusted.
wt	A string "modwt" (maximal overlap discrete wavelet transform) or "dwt" (discrete wavelet transform).
interval	A vector containing the end-points of the interval to be searched for the minimum. See: stats::optimize.
tol	The desired accuracy. See: stats::optimize.

Value

A one-dimension double vector indicating the a prior.

Examples

prior_a(df = gdp[1],
        lags = c(1, 1), .H = 2, .K = 8,
        .diff = TRUE, .var = "ar_out", vscale = "level")

nelson16silva/wavdrcast documentation built on April 25, 2021, 7:03 a.m.