mfp 1.5.5
- Fixed a NOTE (link to included package missing) and removed an
irrelevant argument.
 
mfp 1.5.4
- Fixed a bug where variables with nested names were incorrectly
excluded from the model.
 
mfp 1.5.3
- Added term predictions including first and second derivatives.
 
- Rewrote vignette as an .Rmd file and added a terms prediction
example.
 
- Corrected the use of offsets in glm models.
 
- Changed rescale default: Now, by default `rescale = FALSE’.
 
mfp 1.5.2
- Minor corrections (Vignette float options error; function
namespace).
 
mfp 1.5.1
- Minor corrections (Vignette keywords updated).
 
mfp 1.5.0
- New maintainer; ‘vignettes’ directory added.
 
mfp 1.4.9
- Minor corrections in fp.rescale.
 
mfp 1.4.8
- Call of final fitted glm/coxph model corrected (parent frame was
missing).
 
- plot.mfp (temporarily) removed.
 
- Correction of rescaling procedure of log trafos in fp.rescale.
 
mfp 1.4.7
- summary.mfp added to avoid problems calling summary directly on mfp
object.
 
mfp 1.4.6
mfp 1.4.5
- Minor corrections of component ‘fit’.
 
mfp 1.4.4
mfp 1.4.3
- Corrected in object$fit: Use quadratic in ln(X) when
powers=(0,0).
 
mfp 1.4.2
- Minor corrections to use ‘subset’ properly and to allow to ignore
‘data’.
 
mfp 1.4.0
- Data set ‘GBSG’ from the JRSS-A paper replaced by its Stata
version.
 
- ‘rescale’ argument added to function call.
 
- ‘keep’ argument added to function call.
 
- Verbose output now shows also transformation functions of
covariates.
 
mfp 1.3.2
- Scaling of variables is added to model fit used for output component
‘fit’.
 
mfp 1.3.1
- Output corrected for FP2 with equal powers.
 
- mfp now works even without any FP call.
 
- New argument ‘method’ to select method for tie handling in Cox
models.
 
- Vignette updated.
 
mfp 1.3.0
- Changes in the primary ordering of variables.
 
- Scaling changed.
 
- Likelihood ratio tests are now used for all selections.
 
- Plot of functional form corrected.
 
- Print function corrected.
 
- Value of mfp fit has new component `fit’.
 
mfp 1.2.2
- plot.mfp changed: new argument `var’.
 
- New dataset `bodyfat’ to be used as an example for linear
regression.
 
- Vignette included.
 
mfp 1.2.1