Corrects some minor issues in the manual.
Add the function porthogonalize.list to orthogonalize a sequence of
projectors. It externalizes the process for othogonalizing projectors previously performed within pstructure.formula.
Add the function mat.ginv for computing generalized inverses.
Add the functions mat.cor and mat.corg for forming correlation matrices.
Fix CRAN errors and increment version number.
Correct the DesignNotes vignette.
facetstrips.switch argument to designGGPlot and the
facetstrips.placement argument to designGGPlot and designBlocksGGPlot.
Correct the DesignNotes vignette.
Add a check for a term that is equivalent to the grand mean term in designAnatomy.
Revise the DesignNotes vignette.
Update ggplot2 functions to use tidy evaluation.
replications argument to designAmeasures and implement the calculation of weighted average pairwise variances using the
Rebuild for R version 4.1.2.
Fix bug in the function fac.multinested that resulted in the order of the levels for each new nested factor, when each has the same levels, depending on the order that they occurred within the level of the
Add the function fac.uncombine.
Add the functions fac.split and fac.genfactors.
Modify fac.gen so that the number of levels for each recipient factor is the number of observed levels of that factor.
Fix bug in finding marginal terms in pstructure.formula when there is partial aliasing amongst terms.
Add the function fac.recast, a function that replaces and extends fac.recode.
Fix bug in pstructure that resulted in errors in aliasing reporting.
scales.free argument to allow the axis scales to vary between facets in designGGPlot.
Recompile for submission to CRAN
Fix bug in pstructure for parenthesizing terms that consist of generalized or combined factors.
Fix bug in the aliasing table, where "## Information remaining" not printed for pstructure.
... argument to print.aliasing call in print.structure.
Allow multiple values for
which argument in print.pstructure.
Minor refinements of the the tables of aliasing generated by pstructure.formula, projs.2canon and designAnatomy.
Tidy up print.pstructure by removing the ineffective
which.criteria option and adding a
which option to control which components are printed.
Change the tables of aliasing generated by pstructure.formula, projs.2canon and designAnatomy so that the amount of information that is confounded is reported, instead of the amount of information that remains. The amount that remains is reported in the decomposition table.
Rebuild vignettes and and package for R 4.0.3 and CRAN submission.
Fix a bug in print.summary.pcanon that resulted in incorrect consecutive repetitions of sources within a column.
Check that at least one of
cellfillcolour.column is set in a call to designGGPlot.
labeller arguments to designGGPlot.
Re-implement fac.nested so that it is much more efficient.
Rebuild vignettes and and package for R 4.0.2 and CRAN submission.
x.axis.position argument to designGGPlot, making
top the default.
Rebuild for R version 4.0.0.
Minor revisions to
Change the Maintainer email address.
Add function mat.vcomp.
Add the function rep.data.frame.
Rename the design species for which designTwoPhaseAnatomies produces anatomies. Also, allow the overriding of the default titles for each anatomy.
Add the function fac.uselogical.
Reduce example for fac.multinested and build for CRAN resubmission.
Rebuild for CRAN submission.
Add names of countries to Countries factor in Casuarina.dat.
Make additions to the DesignNotes vignette.
Add names of missing factor/covariate to error message in pstructure.
Add function fac.multinested
Revise the DesignNotes vignette.
Further debug designRandomize for some
data.frames not in standard order.
Correct typo and re-upload to CRAN.
Prepare for upload to CRAN.
Add the capability to mat.Vpredicts to return either the variance or the information matrix. Also, add the rank of the matrix as an attribute of the matrix.
Revise DesignNotes vignette.
Further debug designRandomize for either
data.frame not in standard order or column order inconsistencies.
Rename name of data set
Include check for names supplied in
nested.recipients are in the names in
Fix bug in designRandomize that occurs when a supplied
data.frame is not in standard order.
Update the DesignNotes vignette.
Merge features of designPlot and designPlotlabels into a new function designGGPlot and replace blockboundaryPlot with designBlocksGGPlot. The functions designPlot, blockboundaryPlot and designPlotlabels have been marked for future deprecation.
Update the DesignNotes vignette.
Rebuild under version 3.6.0 prior to resubmission to CRAN.
Update the Author field to include an ORCID and the Description field of the DESCRIPTION.
Add new function designTwophaseAnatomies.
reverse.y arguments to designPlotlabels.
Fix bug in mat.Vpredicts when there are no random effects.
Use R.rsp to install vignettes so that hyperlinks are active.
Deal with a mean only formula (~1) in pstructure.formula.
Reorganize vignette and include manual as a vignette.
Minor updates to the daeDesignNotes vignette and move to a vignettes subdirectory.
Add diagnostic message to mat.Vpredicts.
Fix a bug in setValidity of class "projector".
Fix bug in mat.Vpredicts when have a a single random term for a single factor.
Add the new function mat.Vpredicts that has renamed arguments and the capability of using formulae in generating the required matrices.
eliminate argument to mat.Vpred.
Fix bug in summary.pcanon for a single
projectors are saved.
Small documentation edits.
Rebuild under latest version of R 3.5.0 with requirement for R >= 3.1.0.
Rebuild under latest version of R 3.6.0.
Fix bugs in designAnatomy when there is only one structure.
Fix bugs in print.summary.pcanon when there are no entries in a column of the table.
Add a warning to print.summary.canon when the sum of sources' dimensions is less than the number of rows in data.
Modified checking of options for function arguments.
Modified mat.Vpred so that the Moore-Penrose inverse, which is unique, is always used for matrix inverses.
Rebuild under latest version of R.
Minor editing of documentation.
Changed designRandomize so that, if
NULL, then a random permutation of the recipient factors is produced.
Adjust caption of summary.pcanon.
Added new function designPlotlabels that uses ggplot to plot labels on a grid.
Correct bug in axis labelling in interaction.ABC.plot.
Reinstate ability to form and print a summary from old pcannon objects (pre version 3.0-0).
Fixed a bug in summary.pcanon so that aliasing table is printed when required.
Put in a check that factors/covariates in a formula are in data.
Formatted the aliasing table to be similar to the decomposition table.
Apply David Stern's fix to fac.nested so that it deals with missing values in the nesting factor.
Added links to documentation.
Added a component named
pstructure.object that stores the information about (partial) aliasing between sources derived a single
formula. Added a component named
pcanon.object that collects all the aliasing that has occured in the decomposition.
Added the function print.aliasing to print a data.frame of class
aliasing that contains aliasing information.
Renamed Ameasures to designAmeasures
Replaced the vignette for randomzing designs with a vignette that discusses the facilities for designing experiments that includes the use of designRandomize and designAnatomy.
Added the new function designRandomize with the intention that it will eventaully replace fac.layout. The new function replaces the arguments
allocated, respectively, and
nested.recipients. Also, it changes the default for
Deprecated projs.structure, replacing it with pstructure.formula, creating a pstructure.object that stores the projectors, the lists of their term names and source names, and the marginality matrix for the terms. Added arguments
check.marginality to allow the marginality matrix to be supplied, overwriting any that is produced by pstructure.formula. The argument
check.marginality allows one to decide if the supplied and calculated marginality matrices are the same. Methods for as.data.frame, marginality and print were added for objects of class
Deprecated projs.canon, it being replaced by designAnatomy.
Modified the function designAnatomy to accomodate the changes to pstructure. Added new argument
labels to designAnatomy, which has option
sources that will result in sources indicating whether they represent interactions and/or nested effects. Added argument
marginality that allows a list of marginality matrices to be supplied. Added the argument
check.marginality allows one to decide if the supplied and calculated marginality matrices are the same.
labelswap to summary.pcanon so that can swap between terms and sources in the output.
Added new method marginality.pcannon to extract the list of marginality matrices corresponding to the
formulae in the call to designAnatomy. Added the argument
labelswap to the summary.pcanon method so that either terms or sources can be output.
Added the oats split-plot design from Yates (1937).
Converted projs.canon to designAnatomy, changing the position of
data to be second in the argument list and changing
grandMean. Retained projs.canon as a synonym with the same arguments as before. Also changed
grandMean in projs.structure.
Added new functions mat.ncssvar and Zspline that calculate the varaince matrix of the random spline effects and the design matrix for fitting random spline effects, respectively.
In designPlot, have the whole
designMatrix plotted if
Add the more general
hybrid option to projs.structure and projs.canon and made it the default option.
Changed fac.combine so that a single factor would pass through without generating an error message.
Revise interaction.ABC.plot to account for ggplot deprecation.
Add functions mat.sar, to form the correlation matrix from an SAR parameter, and mat.gau, to form a correlation function based on the square of the distance between coordinates.
Fix a bug in tukey.1df when there is no Error function in the model.
Fix bug in no.reps when the number of reps to achieve the desired power is less than 2.
Fix bug in fac.combine that can introduce a level equal to
NA when none of the factors has such a level.
Rebuild under R 3.4.0.
Include the argument
meanTerm in projs.structure and projs.canon to allow the option to include the grand mean term.
Fixed a bug for aliased terms in projs.jandw.
Extended Ameasures to include a
groups argument that allows explicit specification of the elements of groups for which A values are required.
Fixed a bug in fac.gen when there is a factor with a single character value.
Revised fac.recode and fac.divide to deal with factors with both numeric and character levels and to improve efficiency.
Revised some documentation.
Added functions mat.Vpred and Ameasures.
Fixed bugs in fac.vcmat.
Reversed order of parameters in mat.ar1 and mat.exp.
Added functions mat.ar2, mat.ar3, mat.ma1, mat.ma2, mat.arma and mat.sar2.
xefficiency, the maximum efficiency factor projector-based functions.
Improved the code for analyzing the relationships between pairs of projectors.
Include new functions designPlot and blockboundaryPlot, that are the same as design.plot and blockboundary.plot, except that they have less cryptic argument names and some arguments have been modified. It is envisaged that the new functions will supersede the old functions with the old functions being deprecated in due course.
Fixed bug in projs.structure to ensure that arguments, including
keep.order, are passed to the terms function.
Add arguments rchtdiv and cchtdiv to design.plot so that can size row and column labels independetly of the cell labels.
Improved the documentation of efficiency criteria in a number of functions.
Allow a single formula to be specified in formula of projs.canon.
Improved proj2.combine to equate Qconf to Q1 when all of Q1 is confounded by Q2.
Fix bug in mat.banded in which the elements were being placed starting with the first subdiagonal instead of the diagonal.
ggplotFuncs to interaction.ABC.plot to allow
ggplot functions to be used to
modify the plot.
Add detect.diff to allow the calculation of the detectable difference for given replication, power, delta and alpha.
differencing in projs.structure and projs.canon.
Code modified so that differencing only orthogonalizes by subtracting from a projector those
projectors whose factors are a subset of its factors.
Added new criteria
dforth to efficiency.criteria.
order to the default setting of
which.criteria in all functions that have this argument.
documentation of deprecated functions added.
An except argument added to fac.layout so that nominated unrandomized factors are not permuted.
Added functions to perform an eigenalysis of the projection matrices for an experimental design to produce the decomposition or skeleton-anova table. These functions include projs.combine.p2canon, projs.structure, projs.2canon, projs.canon, summary.p2canon, summary.pcanon, print.summary.p2canon, print.summary.pcanon. The top-level functions are those with the suffix 'pcanon'.
This version uploaded to CRAN.