Now the lower triangular expansion on the matrix of loadings is optional:
lowerTriangular: logical value indicating whether a lower triangular parameterization should be imposed on the matrix of factor loadings (if TRUE) or not. Default: TRUE.
Now the lower triangular expansion on the matrix of loadings is optional.
In this new version I have removed all dependence on the orphaned doRNG package.
Pre-print available on arXiv: Clustering Multivariate Data using Factor Analytic Bayesian Mixtures with an Unknown Number of Components
Bug fix in plot
for Lamba_map
.
Bug fix in plot
for Lamba_map
.
Improved plot and summary methods. See CRAN page.
Improved plot
and summary
methods. This new version will be available to CRAN later this month.
This version fixes a bug to output which caused the main function to crash at certain cases. Not on CRAN yet.
See CRAN page
The fabMix
function now features a new argument (parallelModels
), allowing the user to run different models in parallel. This is combined with the pre-existing option to run heated chains in parallel, thus, parallelization is now implemented in both model-level and chain-level.
[September 2018] Versions 3 and 4 added, which contain a plethora of new models and functionalities.
Version 3.0 introduces new parsimonious models.
Since version 4.0, the package is integrated with C++ code.
Version 4.1 improves the Lambda output: all MCMC values are now exported to a single file (instead of multiple ones, as done in previous versions).
[March 2018] Paper accepted to Computational Statistics & Data Analysis
Please use version 2.0 of the software for the relevant R scripts. But note that
- Recent versions are recommended.
- All package versions smaller than 4.2 are compatible only with Linux distributions.