# -*- org -*-

* Version 0.0.16
the totL2 now also works on matrices with missing data by temporarily setting NA to zero.
* Version 0.0.15
** DONE Error with method 'one'
   SCHEDULED: <2010-05-14 Fri> CLOSED: [2010-05-14 Fri 11:18]

should have worked but croaked:

: Found missing or negative values - applying imputation using ppca
: Error in solve.default(CtC) : 
:   system is computationally singular: reciprocal condition number = 0
: In addition: Warning message:
: In pca(lsta, method = "ppca") :
:   more components than matrix columns requested
: Error in completeObs(pca(lsta, method = "ppca")) : 
:   error in evaluating the argument 'object' in selecting a method for : function 'completeObs'

[[file:R/norm.R::one][fixed]], dont do pca on single column data and never for method one.

* Version 0.0.14 <2010-01-14 Thu>
- Changed the warning about imputing missing values to a message
  instead as imputation not necessarily is anything that needs the
  user's attention.
- More adaptations for using both pcaMethods 1.24.0 and 1.27.2
- The column name in fData and what it should equal for ISs can now be
  set the user. Asked by Pernilla Lindén.
* Version 0.0.12 <2010-01-08 Fri>
- Bugfixes and adapations to make crmn work with pre 1.26.0 versions
  of pcaMethods.
* Version 0.0.10 <2010-01-04 Mon>
- The pcaMethods package now contain Krzanowski style cross-validation
  which yields more stable results and crmn now use this CV by
  default.