The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
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[@]dogado.de.
mspca() and tpm() now take two possible
inputs: the covariance/correlation matrix or the data matrix directly.
In practice, the functions take single generic argument M
together with a type = c("Sigma", "X") selector.
type = "Sigma" (the default) treats M as a
covariance/correlation matrix (p x p); type = "X" treats
M as a raw data matrix (n observations x p variables). The
"Sigma" default preserves the behaviour of existing
matrix-based calls.Sigma %*% beta is evaluated as
t(X) %*% (X %*% beta) / (n - 1) at cost O(np), and the p x
p matrix is never materialized. This substantially improves scalability
when n << p. The covariance back-end was refactored
behind a covariance-operator abstraction (DenseOp /
GramOp) shared by both input modes.type = "X":
center, scale (covariance vs correlation), and
divisor (“n-1” or “n”).Sigma input is
checked for squareness, symmetry and positive semidefiniteness
(checkPSD, symTolerance,
psdTolerance); an X input is checked for
finiteness, dimensions and (when scaling) zero-variance columns.mspca() results now include
variance_explained (per-PC) and
total_variance; X-mode results also record
inputType, center, scale,
divisor, nObs and p.mspca() and tpm() now return S3 objects of
class "mspca" and "tpm" respectively, enabling
use of standard R generics.print.mspca(): S3 print method displaying the
sparse loading matrix restricted to the union of active variables, the
percentage of variance explained per PC, and the number of non-zero
loadings. Replaces the removed print_mspca().summary.mspca(): produces a per-PC table of
sparsity, variance explained, FVE, and cumulative FVE, followed by the
full pairwise feasibility violation matrix.library(datasets) calls from
examples while keeping explicit datasets::mtcars usage, and
added datasets to Suggests to align example
dependencies with CRAN guidance.pairwise_correlation() and
orthogonality_violation() and replaced them with a unified
feasibility_violation_off() helper for feasibility
diagnostics across constraint types.mspca() and tpw()
through algorithmic and implementation optimizations.mspca() now accepts a new hyper-parameter
minRestartTPM that limits the number of random restarts for
the truncated power method after the first outer iterationfeasibilityConstraintType parameter to
msPCA().orthogonality_violation to
feasibility_violation to support both constraint
types.feasibility_violation() as
orthogonality_violation() to be more explicitpairwise_correlation()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.
Health stats visible at Monitor.