Variable Banding of Large Precision Matrices
Generate an autoregressive model.
Generate a model with block-diagonal structure
Plot the sparsity pattern of a square matrix
Generate random samples.
Compute the varband estimate for a fixed tuning parameter value with d...
Perform nfolds-cross validation
Generate a model with variable bandwidth.
Solve main optimization problem along a path of lambda
Implementation of the variable banding procedure for modeling local dependence and estimating precision matrices that is introduced in Yu & Bien (2016) and is available at <https://arxiv.org/abs/1604.07451>.