Seed control implemented for the functions desla and HDLP. Functionality remains the same, and seeds can be set the standard way with set.seed(). This is only relevant when using the (default) plug-in tuning parameter selection method.
Various fixes, including a bug where the wrong local projections were estimated under certain choices of optional arguments.
The function HDLP_state_dependent is removed, and its functionality incorporated as optional arguments into the function HDLP.
Added a plot function which creates plots of impulse responses based on the objects created by the HDLP function.
The desla function is now parallelized over the nodewise regressions, and displays a progress bar.
Various fixes to prevent crashes.
Added the function HDLP_state_dependent, which allows for estimation of impulse responses in different states, and also an option to estimate them with OLS rather than desparsified lasso, using the same long-run variance estimator.
Fixed bug in the long run covariance function, which occasionally caused crashes