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.

glemb 0.1.4

• Added the nlsymiss dataset (581 children from the NLSY with missing values on five variables), used in all examples.
• Examples now run as-is on nlsymiss instead of being wrapped in \dontrun{}.
• Corrected the Little & Schluchter citation to 1985 and added DOI/ISBN links to the references in DESCRIPTION and the help pages.

glemb 0.1.3

• Fixed a reproducibility bug where calling glemb() with an odd number of imputations (m) caused results to alternate between two distinct sets of imputations across successive calls with the same seed. The imputation step now uses R’s own random number generator instead of the Fortran RNG in the mix package, whose internal state could not be fully reset between calls.

glemb 0.1.2

• Added max.cells argument to glemb() (default 500000). A pre-flight check now stops with an informative error before glemb allocates any memory when the nominal variables would produce more cells than this limit. Set max.cells = Inf to disable the check.
• Added meanmodel argument to glemb(). "main" (default) constrains the continuous variable means to depend only on the main effects of the categorical variables, reducing the number of mean parameters from n_cells to 1 + sum(K_j - 1). "saturated" restores the original cell-specific intercepts (unrestricted means).
• The meanmodel value is stored in the returned object and displayed by print() and summary().

glemb 0.1.1

• Added output = "mitml" option, returning a mitml.list object compatible with mitml::testEstimates().
• Several bug fixes.

glemb 0.1.0

Initial release.

• glemb(): multiple imputation for mixed continuous and categorical data using the General Location Model with the EMB algorithm.
• Factor variables automatically treated as categorical; all other variables treated as continuous.
• Log-linear model restricted to 2-way interactions by default; 3-way interactions available via cat.interact = 3.
• Dirichlet prior on cell probabilities (cat.prior) for sparse-cell stabilisation.
• Ridge prior on continuous covariance matrix (empri).
• Output as list of completed data frames (default) or mids object compatible with mice::pool().
• print() and summary() methods for glemb objects.

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.