NEWS.md

In Tenenhaus/RGCCA: Regularized and Sparse Generalized Canonical Correlation Analysis for Multiblock Data

RGCCA 3.0.2

• Rename RGCCA classes and prefix them with "rgcca_".

• Merge documentation of print functions.

• Merge documentation of plot functions.

• Environment variable TEST_SNAPSHOTS must be set to "TRUE" for snapshots to be evaluated.

RGCCA 3.0.0

• Added a NEWS.md file to track changes to the package.

Many differences have been introduced since the last version published on CRAN. We list below the most important ones: The function rgcca is now the main entry point of the package. Many well-known methods of the multiblock literature are now directly available by setting the method argument in the rgcca function. The list of methods can be found using the available_methods function. Arguments A and C of the rgcca function are now deprecated. blocks and connection should be used instead. Small utility functions are no longer exported. Functions rgcca_transform and rgcca_predict have been added. They allow applying fitted RGCCA models to project new subjects onto the learned subspaces and make predictions of response blocks using caret predictive models. Functions rgcca_cv and rgcca_permutation have been added. They allow for finding the best parameters of the models based on a cross-validation or permutation criterion. The function rgcca_bootstrap has been added to evaluate the weights of a model using a bootstrap procedure. The function rgcca_stability has been added to evaluate the stability of the variable selection performed by SGCCA. * Print and plot functions have been added for the outputs of functions rgcca, rgcca_cv, rgcca_permutation, rgcca_bootstrap, and rgcca_stability.

Tenenhaus/RGCCA documentation built on July 20, 2024, 2:14 p.m.