dagitty v.0.3-2 ============== Changes: * New export options for the causal.effect package and for Python-style edge lists * DAGs can now be specified without the enclosing "dag { ... }" if the specification contains at least one directed or bi-directed arrow. dagitty v0.3 (Release date: 2020-06-11) ============== Changes: * toString: added option 'singular' and 'bnlearn' * toMAG: added function to create a MAG from a given DAG * impliedConditionalIndependencies: adds option 'all.pairs' for 'type' argument * instrumentalVariables: added new algorithm (van der Zander et al., 2015) * simulateLogistic: added function to simulate binary data * localTests: added option 'cis.chisq' for argument 'type' to test categorical variables * isCollider: added function to check for colliders in structure * adjustmentSets: added argument 'max.results' * simulateSEM: added options 'empirical' and 'verbose' * plot.dagitty: show coefficient values for edges. * plot.dagitty: fixed bug where bent undirected lines where not drawn thick but non-bent undirected lines were dagitty v0.2-2 (Release date: 2016-08-26) ============== Changes: * Added capability to specify arbitrary path coefficients for simulateSEM * Semi-parametric conditional independence testing for additive noise using loess smoothing * Plot function now sets margins properly and gives better error messages * Bugfix in adjustmentSets with type="all" when exposure was defined in graph * Bugfix in dconnected and isAdjustmentSet when graphs contain inducing paths with 3 or more nodes * Improved parser for larger graphs