ridgeP
output.CITATION
fileREADME
fileGGMpathStats
:
Incorrectly stated before that all igraph layouts were supported.
Now they indeed are supported.GGMpathStats
now has additional visualization options: It can handle all
layout functions supported by igraph. Moreover, it is now possible to
specify custom coordinates for node-placement.CNplot
function again updated: higher max. iterations for Lanczos methodoptPenalty.LOOCV
is deprecated. Please use optPenalty.kCV
insteadoptPenalty.LOOCVauto
is deprecated. Please use optPenalty.kCVauto
insteadCITATION
fileREADME
filesparsify
now has an additional thresholding option: 'connected'CITATION
fileREADME
fileUgraph
:conditionNumberPlot
is deprecated. Please use CNplot
insteadCNplot
function (above and beyond conditionNumberPlot
):digitLoss
and rlDist
arguments have been removedIaids
Iaids = TRUE
amends the basic condition number plot with interpretational aidsmain
is now a character argumentvalue
now by default takes the value 1e-100 (convenient)edgeHeat
now has aligned x-axis labelsoptPenalty.LOOCV
and optPenalty.aLOOCV
functions
now will no longer produce horizontal and/or vertical lines that fall outside the boundaries
of the figureoptPenalty.LOOCV
now uses log-equidistant penalty grid for optimal penalty
parameter determination (this also enhances the visualization)optPenalty.aLOOCV
function:Ugraph
function:ridgePathS
has been sped up by killing redundant inversioncovML
function has been amended with an argument that indicates if a correlation matrix
(instead of an ML estimate of a covariance matrix) is desired. This offers more flexibility. One can now get the
ML estimate of the covariance matrix, the ML estimate of the covariance matrix on standardized data, as well as
the correlation matrixoptPenalty.LOOCVauto
function has been amended with an argument that indicates if the
evaluation of the LOOCV score should be performed on the correlation scaleoptPenalty.LOOCV
function has been amended with an argument that indicates if the
evaluation of the LOOCV score should be performed on the correlation scaleoptPenalty.aLOOCV
function has been amended with an argument that indicates if the
evaluation of the approximate LOOCV score should be performed on the correlation scaleNEWS
file!CITATION
fileREADME
fileC++
. The package should now be at least
two orders of magnitude faster in most cases..fused
.
Functions tied to (or added with) this module are:isSymmetricPD
isSymmetricPSD
is.Xlist
default.target.fused
createS
getKEGGPathway
kegg.target
pooledS
pooledP
KLdiv.fused
ridgeP.fused
optPenalty.fused.grid
print.optPenaltyFusedGrid
plot.optPenaltyFusedGrid
optPenalty.fused.auto
optPenalty.fused
default.penalty
fused.test
print.ptest
summary.ptest
hist.ptest
plot.ptest
sparsify.fused
GGMnetworkStats.fused
GGMpathStats.fused
covMLknown
GGMmutualInfo
GGMpathstats
:sparsify
: Now always returns symmetric objectsverticle
as used in various functions has been renamed to
vertical
. Sorry for any inconvenience.ridgeS
replaced by the faster C++-dependent counterpart
ridgeP
conditionNumberPlot
function:ridgeS
is deprecated. Please use ridgeP
instead.pathContribution
for usage in
GGMpathStats
function.path2string
for usage in
GGMpathStats
function.pathAndStats
for usage in
GGMpathStats
function.cvl
for usage in
optPenalty.LOOCVauto
function.lambdaNullDist
for usage in
GGMblockNullPenalty
function.blockTestStat
for usage in
GGMblockTest
functionGGMpathStats
optPenalty.LOOCVauto
GGMblockNullPenalty
GGMblockTest
fullMontyS
evaluateSfit
function.
The dir
argument was not properly used previously.optPenalty.aLOOCV
function:for
loops.optPenalty.LOOCV
function:default.target
function:type = DIAES
. Gives diagonal
matrix with inverse of average of eigenvalues of S as entries.GGMnetworkStats
function:ridgePathS
function:plotType = "pcor"
.sparsify
function:top
: retainment of top elements
based on absolute partial correlation.output = "light"
, only the (matrix)
positions of the zero and non-zero elements are returned..ridgeSi
for usage in
conditionNumberPlot
function..eigShrink
for usage in (a.o.)
ridgeS
function.GGMnetworkStats
evaluateSfit
ridgePathS
default.target
evaluateS
function:evaluateS
function is now alignedconditionNumber
function now called conditionNumberPlot
.
The updated function has new features:sparsify
function:type = c("threshold", "localFDR")
to
threshold = c("absValue", "localFDR")
(clarifying nomenclature)threshold
to absValueCut
(clarifying nomenclature)ridgeS
function:optPenaltyCV
function now called optPenalty.LOOCV
, for sake
of (naming) consistency. The updated function has new features:targetScale
option has been removedlog
in optional graph by ln
optPenalty.aLOOCV
function:log
in optional graph by ln
conditionNumberPlot
,
optPenalty.aLOOCV
and optPenalty.LOOCV
functions sped up
considerably for rotation equivariant alternative estimator.
By usage new ridgeS and avoidance of redundant eigendecompositionsridgeS
, conditionNumberPlot
,
optPenalty.aLOOCV
and optPenalty.LOOCV
now \code{"DAIE"
option from default.target
covML
optPenalty.aLOOCV
conditionNumber
to visualize the
spectral condition number over the regularization pathevaluateS
to evaluate basic
properties of a covariance matrixKLdiv
that calculates the
Kullback-Leibler divergence between two normal distributionssparsify
functionoptPenaltyCV
functionoptPenaltyCV
and optPenalty.aLOOCV
now utilize
covML
instead of cov
optPenaltyCV
(as in
optPenalty.aLOOCV
) is now light
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.