- Bugfix in learn.network when providing an initial CPC
- Update test files
- Update igraph function calls
- Edited citation file, documentation and some internal function names to comply with CRAN notifications
- Vignette is now static to fix CRAN NOTE/Warning
- Edit vignette to silence CRAN NOTE/Warning
- Bugfix in belief propagation when processing the observed variables
- fix -Wpointer-sign warnings in c code for CRAN
- dag.to.cpdag() now takes into account layer.struct (thanks to Erica Tavazzi)
- class(x)==y ported to inherits(x,y)
- bugfix in BP in case of disconnected single nodes
- Removed direct dependency from Matrix package
- Bugfix in BP (found by Fritz Bayer)
- Bugfix in EM, continuous case
- Speedup EM by computing log-likelihood only for observations that change
- Bugfix in EM
- Improved robustness in reading data and header files
- Bugfix in EM
- Bugfix in SM
- Bugfix in EM algorithm (found by Anne Helby Petersen)
- Correction in vignette example
- Regenerated example data, now it requires R 3.5+
- Added qgraph option for plot()
- Parameters wm.max, initial.cpc exposed in learn.network
- Layering added to HC
- Mandatory edges can be specified to learn.network
- Modified way of imposing a maximum number of parent nodes for the different algorithms
- Bugfix in discretization (thanks to Amy Willis)
- Second revision for paper submission
- First revision for paper submission
- First CRAN submission