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.
error_on = "warning" to detect either “out-of-bounds” or
“nan is outside the range” issuesas_sem(.) because sem::sem(.)
appears to cause a UBSAN warning, and removing sem from
IMPORTSgmrf_parameterization = "gmrf_project", given
V = t(G)*G, to invert Vinv = invertSparseMatrix(V) where Vinv is then
dense and then casting Vinv2 = asSparseMatrix( Vinv ), rather than a
sparseLDLT for solve( V, I-P ), because the latter seems numerically
unstable when P has a high condition number (e.g., the moose-wolf
vignette in tinyVAST)gmrf_parameterization = "conditional_krig" as
gmrf_parameterization = "mvn_project" and confirmed it in
simple casegmrf_parameterization = "separable" as
full and projection as
projectpredict for type="link"
was pulling x_tj rather than z_tj and
therefore was missing the initial conditions, mean value, and only
worked for gmrf_parameterization = "separable"stabilize_Q option to dsem_control,
adding a diagonal component to t(Gamma)*Gamma to ensure it’s PDgmrf_parameterization = "gmrf_project" (which
maintains sparsity while allowing latent variables with zero exogenous
variance, or manifest variables with measurement error and zero
exogenous variance), and which automatically switches to
separable if no variables have zero variancegmrf_project,
mvn_project and separable (with a small extra
non-zero variance inflation) are all identical in a logistic regression
that involves loops (i.e., zero-variance for intermediate latent
variable)dsem_control to use
gmrf_parameterization = "gmrf_project" as defaultdsemRTMB (which threw
an error with updates to RTMB, and is not being used anyway)NA/NaN function evaluation function evaluations from
nlminb (enabled by default, but overriden in
dsem_control)Matrix::bandSparse to simplify logic in
make_matricespartition_variancetotal_effect to compute result for either pulse
or press experimentgmrf_parameterization = "conditional_krig" that excludes
family = "fixed" variables from the GMRF density
calculation while still conditioning upon their value(s)make_matricesggm from Imports (because it Imports
graph which is not on CRAN) and instead define a locally
copy of relevant functions. Also adding Giovanni M. Marchetti (the
maintainer of ggm as contributor in DESCRIPTION)test_dsep(.) to allow options for imputing
missing data prior to test, and using default
impute_data = "by_test" based on explorations to datetotal_effecttest_dsep(.) to have option to impute missing
data, and changing that to be the default behaviortest_dsep(.) to calculate a p-value for the
probability that a a data set arose from the specified model (highly
experimental)total_effect(.) to calculate total effectsquiet = FALSE and
running without data e.g., as qualitative network modelstepwise_selection for automated stepwise model
selectionplot option to use ggraph as
alternative to previous igraph optionconvert_equations to extend
sem::specifyEquations and simplify specification for large
modelsprior_negloglike as interface to
specify Bayesian priors and/or likelihood penalties in
dsemdsemRTMB using RTMB as interchangeable
alternative to dsem using TMBcheckDepPackageVersion(dep_pkg="Matrix", this_pkg="TMB")
from .onLoad() as requested by K. Kristensencovstsdata had two or more columns
sharing a single variable namemake_dfa helper functioneval usagetsdata
have a standard deviation by defaultsimulate.dsem to keep up with changing
interface in dsemThese 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.