Estimation of high-dimensional dynamic conditional precision matrices with an application to forecast combination

The estimation of a large covariance matrix is challenging when the dimension p relative to sample size n. Common approaches deal with challenge have been based on thresholding or shrinkage methods in estimating matrices. However, many applications (e.g., regression, forecast combination, portfolio selection), what we need not but its inverse (the precision matrix). In this paper introduce method high-dimensional “dynamic conditional precision” (DCP) proposed DCP algorithm estimator unconditional high-dimension and dynamic correlation (DCC) model embed structure matrix. simulation results show that performs substantially better than matrices methods. Finally, examine “forecast combination puzzle” using DCP, thresholding,