Rboot {slm}

Risk estimation for a tapered covariance matrix estimator via bootstrap method

Description

This function computes an estimation of the risk for the tapered covariance matrix estimator of a process via a bootstrap method, for a specified treshold and a specified kernel.

Usage

Rboot(epsilon, treshold, block_size, block_n, model_max, kernel_fonc)

Arguments

epsilon

numeric vector. An univariate process.

treshold

integer. Number of estimated autocovariance terms that we consider for the estimation of the covariance matrix.

block_size

integer. The size of the bootstrap blocks. block_size must be greater than model_max.

block_n

integer. Blocks number used for the bootstrap.

model_max

integer. The maximal dimension, that is the maximal number of terms available to estimate the covariance matrix.

kernel_fonc

function. The kernel to use. The user can define his own kernel and put it in the argument.

Value

This function returns a list with:

risk

for one treshold, the value of the estimated risk.

SE

the standard-error due to the bootstrap.

References

E. Caron, J. Dedecker and B. Michel (2019). Linear regression with stationary errors: the R package slm. arXiv preprint arXiv:1906.06583. https://arxiv.org/abs/1906.06583.