The Center for Causal Discovery has released the newest version of its causal discovery software based on Tetrad (Version 6.7). Associated command-line, Python and R implementations also inherit algorithm updates. Please note that previous saved sessions will not load in this new version.
What’s New in Tetrad
Can override variable datatypes using metafile Functions for handling experimental variables Updated default parameters based on empirical simulation Extended parameter descriptions (rollovers) Updated manual Experimental algorithms are now only available in experimental mode i.e., –experimental Model scores (e.g., BIC) are output where available New look for workflow arrows Bootstrapping now includes original dataset by default Revised production algorithm list Improved covariance calculation speed and precision Many bug fixes
Changed edge coloration behavior and docs to match original specification Added sortable table view of edges in graph Generalized bootstrapping to sampling with flexible options Layout fixes in UI Added .dot output for graphs Added KCI Added MultiFASK Updated CCI algorithm