NEWS
ClusterSignificance 1.7.3
- Adding warn argument to pcp. Mostly for internal use. Fixes the warning with non default df argument repeating each perumtation.
- Minor plotting improvments including providing a fix for the plot margin error.
ClusterSignificance 1.7.2
- Tidying and fixing bugs in README. README now from README.Rmd to
avoid similar problems in the future.
- Adding codecov badge to Readme
ClusterSignificance 1.7.1
- Updating author info and package URL.
ClusterSignificance 1.5.3
- Fixed bug in pcp 2D projection plot where colors did not match
classification plot.
ClusterSignificance 1.5.2
ClusterSignificance 1.3.2
- added p-value estimate confidence intervals via conf.int function
- Some vignette updates including adding the "common questions" section.
ClusterSignificance 1.3.1
- Added plotting of individual comparisons in classification and permutation
plots.
- Added df argument (degrees of freedom, passed to smooth.spline) to
projection and permute function. This allows some degree of control over
how linear or crooked the principal curve is drawn. NOTE: you must (!)
give the same df value to the projection and permute functions for your
results to be valid and, at the moment, there is no automated checking
that this is the case.
- The classification step now uses a mew method for separation scoring and
fixes a previous bug which could occur when group sizes were not equal.
This change is reflected in the vignette.
- Fixed bug in concatenation c(), when some iterations fail for a specific
comparison thus causing a error when concatenating permute results.
ClusterSignificance 1.2.0
NEW FEATURES
- Added a slot for group colors for the classes Pcp, Mlp, ClassifiedPoints
and PermutationResults
ClusterSignificance 1.0.0
The ClusterSignificance package was born