The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.
| NEWS | R Documentation |
Removed support for graphical models via package QUIC as QUIC is no longer maintained and was archived on CRAN.
Fixed CRAN checks regarding all.equal by adding
check.environment = FALSE.
Added reference to stability selection with gamboostLSS.
Make vignettes conditional on suggested packages (closes #24).
Fixed tests.
Move packages from Enhances to Suggests.
Make manuals work without package TH.data.
Speed up examples by adding \donttest in two occassions.
Use markup \doi for DOIs in manual.
Added DOIs to ‘DESCRIPTION’ to fulfill CRAN policies.
Updated ‘README.md’ and vignettes.
Added support for graphical models via package QUIC (closes #12, #13 and #14; Code contributed by Richard Beare).
Better handling of errors in (single) folds of stabsel:
results of folds without errors are used and a warning is
issued (closes #2
and #7).
Parallel computing via mclapply: Set
mc.preschedule = FALSE per default (closes
#1).
Make sure that per default glmnet.lasso selects at maximum
q variables. The previous anti-conservative version
stays available via args.fitfun = list(type = "anticonservative") (closes
#23)
Updated references: Hofner et al (2015) now available at BMC
Bioinformatics. See citation("stabs") for details.
Updated ‘DESCRIPTION’ to be more informative.
Make fitfuns more quiet (closes
#19;
Code contributed by Gokcen Eraslan)
Replaced base-learner with effect in warnings and
error messages (closes
#10).
Added example on using glmnet.lasso (closes
#4).
Boosting specific changes: warn if mstop is to small.
Disallow specification of penalty parameters via args.fitfun
(closes #22).
Fixed checks for results of fitfuns
(closes #17; Code contributed
by Andrey Tovchigrechko)
Fixed citation. (closes #9).
Fix plot labels if matrices are used (as opposed to data.frames)
(closes #6).
Fix issue when variables are dropped from active set in
lars.lasso (closes #5).
Adhere to CRAN policies regarding import of base packages (closes #3).
Changes in ‘inst/CITATION’ to make CRAN happy: Citations can now be extracted without the need to install the package.
Changed require to requireNamespace
Added a function stabsel.stabsel() to compute (new)
parameter combinations for a fitted stabsel object
Included selected() method (originally from
mboost); selected() can now be used on stability
selection results to extract selected effects
Added functionality to extract parameters from fitted
stabsel objects:
use parameters() or stabsel_parameters()
Export run_stabsel to make the actual “fit”
function accessible for other packages.
Tweaked print methods a bit
Added output that states the significance level.
Added new lasso based fit (glmnet.lasso_maxCoef)
function where the q strongest predictors (according to the
coefficient size) are selected (feature request from Rajen Shah
<R.Shah _at_ statslab.cam.ac.uk>)
Package development moved from R-forge to https://github.com/hofnerb/stabs
Added specifiedPFER to stabsel and
stabsel_parameter objects (needed for better outputs)
Added ‘inst/CITATION’
Added reference to stability selection paper
stabs now enhances gamboostLSS
Made tests conditional on availability of packages
Initial submission to CRAN
stabs implements stability selection for a range of
models, including a novel matrix interface for stabsel
that can be used with generic fit functions.
The code is based on the stabsel function that was
implemented in mboost until version 2.3-0. stabsel is
now a generic function with a specialized method for boosting models
that is implemented in mboost. The back-end is now completely
implemented in package stabs.
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.