odpc 2.0.4
==============
* Adds a function, crit.sparse_odpc, for computing a sparse version of odpc, where the L1 norm of the a vector is penalized.
The number of lags, components and the penalty constant are chosen automatically.
odpc 2.0.3
==============
* Add "gradient" optimization method, that alternates between solving least-squares problems for computing the loadings matrix and doing gradient descent steps for updating the vector of coefficients that defines the dynamic principal component. This is the new default when the dimension of the input matrix is greater than 10 and provides huge (up to ~10x) performance boosts for matrices with a large number of columns.
* Fix bug in forecast.odpcs when add_residuals=TRUE. Thanks to Sebastian Krantz for spotting this.
* Don't make components generic, rename to components_odpcs. Thanks to Sebastian Krantz for spotting this.
odpc 2.0.2
==============
* Fixed CRAN warnings.
odpc 2.0.0
==============
* The higher order ODPC are now defined as linear combinations of the observations (instead of linear combinations of the residuals of the previous fits). This way
a smaller number of observations is lost when new components are added. The forecasting performance of the new definition is very similar to that of the old one.
* Option to use gdpc as starting point for the iterations is no longer supported (for now).
* Added functions to automate the choice of the tuning parameters for odpc: cv.odpc is based on minimizing the cross-validated forecasting error, crit.odpc
is based on minimizing and information criterion.
* Added several tests.