This function computes the estimated covariance matrix of the coefficients on individual signals using the Moving Block Bootstrap approach considered in Zhang et al. (2021). The detailed definition can be found in Equation (8) in Zhang et al. (2021) (https://arxiv.org/abs/2106.15997).
est_cov(Xt, scale_weights, sB = 10, B)A matrix of dimension T by p, where T is the length of the time series and p is the number of processes.
A vector that denotes the weights on scales. All elements should be non-negative and sum to one.
A double that denotes the positive constant C associated with the block size, which is defined as floor(C*T^1/3). Default value is 10.
An integer indicating the number of Monte-Carlo replications used in the Moving Block Bootstrap.
A matrix of the estimated covariance matrix of the coefficients on individual signals.