NEWS file for the partykit package

Changes in Version 1.2-23 (2024-12-02)

Buxfixes

• Make sure weights are propagated properly in data\_party.

Changes in Version 1.2-22 (2024-08-17)

Buxfixes

• Use R\_Calloc.

Changes in Version 1.2-21 (2024-07-18)

Buxfixes

• Rd link problem.

Changes in Version 1.2-20 (2023-04-11)

Buxfixes

• density now gives warnings when called with weights, avoid this.

• Check for tied p-values fails on M1mac.

Changes in Version 1.2-19 (2023-03-19)

Buxfixes

• Fix documentation bug.

Changes in Version 1.2-18 (2023-03-09)

Buxfixes

• Register default methods.

Changes in Version 1.2-17 (2023-02-28)

Misc

• Update CITATION file.

Buxfixes

• Allow NA in responses when plotting. Reported by Tyson H. Holmes.

Changes in Version 1.2-16 (2022-06-20)

Buxfixes

• Address random CRAN errors (honesty checks).

Changes in Version 1.2-15 (2021-08-23)

Small Improvements

• Update reference output.

Changes in Version 1.2-14 (2021-04-22)

Bugfixes

• cforest ignored converged argument.

Changes in Version 1.2-13 (2021-03-03)

Small Improvements

• Suggest randomForest.

• Test constparty vignette code in tests, to avoid a NOTE about missing RWeka on Solaris.

Changes in Version 1.2-12 (2021-02-08)

New Features

• Add method argument to glmtree. The default is to use "glm.fit" (as was hard-coded previously) but this can also be changed, e.g., to "brglmFit" from brglm2 for bias-reduced estimation of generalized linear models.

Bugfixes

• Fix LaTeX problem.

• Better checks for response classes, fixing a bug reported by John Ogawa.

• In lmtree and glmtree the "xlevels" attribute for the regressors is preserved in the models fitted within the trees. Thus, predicting for data whose "xlevels" do not match, an error is generated now (as opposed to warning and partially incorrect predictions).

Changes in Version 1.2-11 (2020-12-09)

New Features

• Add an experimental implementation of honesty.

• Add maxvar argument to ctree_control for restricting the number of split variables to be used in a tree.

Bugfixes

• all.equal must not check environments.

• Non-standard variable names are now handled correctly within extree_data, prompted by https://stackoverflow.com/questions/64660889/ctree-ignores-variables-with-non-syntactic-names.

Changes in Version 1.2-10 (2020-10-12)

Bugfixes

• Deal with non-integer minsize in mob.

• Handle NAs in .get_psplits.

• Fix URLs.

Changes in Version 1.2-9 (2020-07-10)

Bugfixes

• Fix an issue with printing of tied p-values.

Changes in Version 1.2-8 (2020-06-09)

Bugfixes

• pruning of modelparty objects failed to get the fitted slot right.

• In R-devel, c(<factor>) now returns factors, rendering code in .simplify_pred overly pedantic.

Changes in Version 1.2-7 (2020-03-06)

Bugfixes

• NAMESPACE fixes: party is only suggested.

Changes in Version 1.2-6 (2020-01-30)

Bugfixes

• Remove warning about response not being a factor in predict.cforest. Reported by Stephen Milborrow.

Changes in Version 1.2-5 (2019-07-17)

Bugfixes

• Trying to split in a variable where all observations were missing nevertheless produced a split, as reported by Kevin Ummel.

Changes in Version 1.2-4 (2019-05-17)

Bugfixes

• update reference output, fix RNGversion

Changes in Version 1.2-3 (2019-01-28)

New Features

• varimp runs in parallel mode, optionally.

• weights in cforest can now be used to specify a matrix of weights (number of observations times number of trees) to be used for tree induction (this was always possible in party::cforest. This was advertised in the documentation but actually not implemented so far.

Bugfixes

• predict did not pay attention to xlev; this caused problems when empty factor levels were removed prior to tree fitting.

• nodeprune may have got fitted terminal node numbers wrong, spotted by Jason Parker.

Changes in Version 1.2-2 (2018-06-05)

Bugfixes

• In mob() using the cluster argument with a factor variable sometimes lead to NAs in the covariance matrix estimate if empty categories occured in subgroups. The problem had been introduced in version 1.2-0 and has been fixed now.

• Methods for the sctest generic from the strucchange package are now dynamically registered if strucchange is attached. Alternatively, the methods can be called directly using their full names sctest.constparty and sctest.modelparty.

• The prune.modelparty function is now fully exported but it is also registered with the prune generic from rpart.

Changes in Version 1.2-1 (2018-04-20)

New Features

• New scale argument for predict in cforest. For simple regression forests, predicting the conditional mean by nearest neighbor weights with scale = TRUE is now equivalent to the aggregation of means. The unscaled version proposed in <doi:10.1002/sim.1593> can be obtained with scale = FALSE.

Bugfixes

• Bug fix for case weights in mob() in previous version (1.2-0) introduced a bug in the handling of proportionality weights. Both cases are handled correctly now.

• glmtree can now handle caseweights = TRUE correctly for vcov other than the default "opg". Internally, the glm objects are adjusted by correcting the dispersion estimate and the degrees of freedom.

• lookahead did not work in the presence of missing values.

• Calling partykit::ctree did not work when partykit was not attached.

• node_inner now allows to set a different gpar(fontsize = ...) in the inner nodes compared to the overall tree.

• splittest asked for Monte-Carlo p-values, even when the test statistic was used as criterion.

Changes in Version 1.2-0 (2017-12-18)

New Features

• We welcome Heidi Seibold as co-author!

• Internal re-organisation for ctree by means of new extensible tree infrastructure (available in extree_data and extree_fit). Certain parts of the new infrastructure are still experimental. ctree is fully backward compatible.

• Use libcoin for computing linear test statistics and p-values for ctree.

• Use inum for binning (the new nmax argument).

• Quadratic test statistics for splitpoint selection are now available for ctree via ctree_control(splitstat = "quadratic").

• Maximally selected test statistics for variable selection are now available for ctree via ctree_control(splittest = TRUE).

• Missing values can be treated as a separate category, also for splits in numeric variables in ctree via ctree_control(MIA = TRUE).

• Permutation variable importance, including conditional variable importance, was added to partykit.

• New offset argument in ctree.

• New get_paths for computing paths to nodes.

• node_barplot gained a text argument that can be used to draw text labels for the percentages displayed.

• The margins used in plot.party can now also be set by the user.

Bugfixes

• Bug fix in mob() if weights are used and caseweights = TRUE (the default). The statistics for the parameter instability tests were computed incorrectly and consequently the selection of splitting variables and also the stopping criterion were affected/incorrect.

• Avoid log(p) values of -Inf inside mob() by replacing weighted averaging with naive averaging in the response surface regression output in case the p values are below machine precision.

• The as.party method for rpart objects without any splits only returned a naked partynode rather than a full party. This has been corrected now.

• nodeapply did not produce the same results for permutations of ids. Spotted by Heidi Seibold.

• Out-of-bag predictions in predict.cforest were incorrect.

• perm in predict was only considered when newdata was given. Spotted by Heidi Seibold.

• Don't try to search for binary splits in unordered factors with more than 31 levels. This potentially caused an integer overrun in previous versions. party::ctree() uses an approximation for binary split searches in unordered factors; thus, using party might be an alternative.

Changes in Version 1.1-1 (2016-09-20)

• Proper support of quasi-families in glmtree and hence palmtree.

• NA handling by following the majority was potentially incorrect in ctree.

• Minor speed improvements.

• Breaking ties before variable selection was suboptimal for very small log-p-values.

Changes in Version 1.1-0 (2016-07-14)

• Added a new function palmtree that fits partially additive (generalized) linear model trees. These employ model-based recursive partitioning (mob) based on (generalized) linear models with some local (i.e., leaf-specific) and some global (i.e., constant throughout the tree) regression coefficients.

• Splits in ordinal variables are now represented correctly in the (still internal) list.rule method.

• Kaplan-Meier curves in "constparty" trees were plotted incorrectly due to use a wrong scaling of the x-axis. Spotted by Peter Calhoun <calhoun.peter@gmail.com>.

• Use quote(stats::model.frame) instead of as.name("model.frame").

• The as.party methods for rpart and Weka_tree now have a data = TRUE argument so that by default the data is preserved in the party object (instead of an empty model frame).

• The predict method for cforest objects did not work for one-row data frames, fixed now.

• Added rot and just arguments to node_barplot for more fine control of x-axis labeling (e.g., with 45 degree rotation).

Changes in Version 1.0-5 (2016-02-05)

• The partykit package has now been published in Journal of Machine Learning Research, 16, 3905-3909. https://jmlr.org/papers/v16/hothorn15a.html

• Added support for setting background in panel functions.

• The as.list() method for partynode objects erroneously created an object thisnode in the calling environment which is avoided now.

Changes in Version 1.0-4 (2015-09-29)

• Bug fix in plot() method for constparty objects. In the previous partykit version clipping was accidentally also applied to the axes labels.

• For constparty objects plot(..., type = "simple") did not work correctly whereas plot(as.simpleparty(...)) yielded the desired visualization. Now internally as.simpleparty() is called also in the former case.

• The as.simpleparty() method now preserves p-values from constparty objects (if any).

• Added a getCall() method for "party" objects.

• In the predict() method for "lmtree" and "glmtree" objects the offset (if any) was sometimes ignored. It is now always used in the prediction.

Changes in Version 1.0-3 (2015-08-14)

• Import logrank_trafo from coin.

Changes in Version 1.0-2 (2015-07-28)

• nodeprune(..., ids = 1) did not prune the tree to the root node. Fixed now.

• predict.cforest used na.omit instead of na.pass.

• predict.party now features new perm argument for permuting splits in specific variables (useful for computing permutation variable importances).

• NAMESPACE updates.

Changes in Version 1.0-1 (2015-04-07)

• The support for (generalized) linear model trees with just a constant regressor has been improved. Now lmtree(y ~ x1 + x2) is short for lmtree(y ~ 1 | x1 + x2), analogously for glmtree(). Plotting now also works properly in this case.

• The as.party() method for "rpart" objects did not work if one of the partitioning variables was a "character" variable rather than a "factor". A suitable work-around has been added.

• The node_barplot() panel function can now also be used for multivariate responses, e.g., when all responses are numeric and on the same scale.

• The package now also includes a new data set HuntingSpiders which is essentially a copy of the spider data from the package mvpart that is currently archived on CRAN. The documentation has been improved somewhat and is likely to change further to explain how the data has been transformed in De'ath (2002).

• The survival tree example for the GBSG2 data was broken due to the response being (incorrectly) also part of the explanatory variables. Fixed by using the latest Formula package (at least version 1.2-1).

Changes in Version 1.0-0 (2015-02-20)

• Version 1.0-0 published. This version is described in the MLOSS paper accepted for publication by the Journal of Machine Learning Research today.

• The unbiased version of cforest (with replace = FALSE) is now the default (as in party).

• Register all S3 methods in NAMESPACE.

Changes in Version 0.8-4 (2015-01-06)

• Extended mob() interface by a cluster argument. This can be a vector (numeric, integer, factor) with cluster IDs that are then passed on to the 'fit' function (if supported) and used for clustering the covariance matrix in the parameter stability tests. lmtree() and glmtree() hence both gained a cluster argument which is used only for cluster covariances but not for the model estimation (i.e., corresponding to a working independence model).

• Optionally, the parameters' variance-covariance matrix in mob() can now be estimated by the sandwich matrix instead of the default outer-product-of-gradients (OPG) matrix or the information matrix.

• Reimplementation of cforest() available with extended prediction facilities. Both the internal representation and the user interface are still under development are likely to change in future versions.

• Added multicore support to mob(), ctree(), and cforest(). If control argument cores is specified (e.g., cores = 4) then the search for the best variable or split point (often involving numerous model fits in mob() or resampling in ctree()) is carried out using parallel::mclapply() rathern than sequential for() or sapply(). Additionally, other applyfuns can be provided, e.g., using networks of workstations etc.

• Bug fix in mob() that occurred when regressor variables and partitioning variables overlapped and were not sorted in the underlying model frame.

Changes in Version 0.8-3 (2014-12-15)

• mvpart was archived 2014-12-15.

Changes in Version 0.8-2 (2014-09-12)

• Fixed an uninitialized memory issue reported by valgrind.

Changes in Version 0.8-1 (2014-09-08)

• partykit now depends on R version >= 3.1.0 in order to import the depth() generic from the grid package.

• The print methods for party/partynode objects with only a root node was modified. Now, the terminal panel function is also applied if there is only a root node (while previously it was not).

• ctree() now catches sum(weights) <= 1 situations before they lead to an error.

• Code from suggested packages is included by using :: syntax as required by recent R versions.

• Argument ytrafo of ctree() can now be a function which will be updated in every node.

• A small demo briefly illustrating some memory and speed properties has been added. It can be run interactively via demo("memory-speed", package = "partykit").

• Section 3 of the "constparty" vignette now shows how properties of a new tree algorithm can be assessed by using partykit building blocks.

Changes in Version 0.8-0 (2014-03-27)

• Major improved version of partykit. The previously existing functions in the package were tested and enhanced, new functions and extensive vignettes added.

• Extended and improved introductory documentation. The basic classes and class constructors partynode/partysplit/party are introduced in much more detail now in vignette("partykit", package = "partykit").

• The class constparty (inheriting from party) for representing party objects with constant fits in the nodes (along with coercion methods for rpart, J48, etc.) is now described in more detail in the new vignette("constparty", package = "partykit").

• The package now includes a reimplementation of the model-based recursive partitioning algorithm (MOB) using partykit infrastructure. The generic algorithm, the corresponding convenience interfaces lmtree() and glmtree() as well as various illustrations and possible extensions are described in detail in the new vignette("mob", package = "partykit").

• Improved implementation of conditional inference trees (CTree), see the new vignette("ctree", package = "partykit") for details.

• New nodeprune() generic for pruning nodes in all party trees and partynode objects.

• Deal with empty levels in ctree() for teststat = "quad" (bug reported by Wei-Yin Loh <loh_at_stat.wisc.edu>).

• In predict() method for constparty objects, type = "prob" now returns ECDF for numeric responses and type = "response" the (weighted) mean.

• New panel function node_ecdf() for plotting empirical cumulative distribution functions in the terminal leaves of constparty trees.

Changes in Version 0.1-6 (2013-09-03)

• Bug fix in as.party() method for J48 trees with ordered factors.

Changes in Version 0.1-5 (2013-03-22)

• Fix C code problems reported by clang under OS X.

Changes in Version 0.1-4 (2012-06-05)

• Added node_surv() for plotting survival ctrees. Accompanying infrastructure for survival trees was enhanced.

• ctree() now checks for (and does not allow) x >= max(x) splits.

Changes in Version 0.1-3 (2012-01-11)

• Added ipred to the list of suggested packages due to usage of GlaucomaM and GBSG2 data in tests/examples.

Changes in Version 0.1-2 (2011-12-18)

• The node_terminal() panel-generating function is now customizable by a FUN argument that is passed to formatinfo().

• The plot() method for simpleparty object now sets up a formatting function passed to formatinfo(), both in print() and plot().

• Fixed bug in pmmlTreeModel() for processing label IDS in splits when not all levels are present.

• Cleaned up unused variables in C code and partial argument matching in R code.

Changes in Version 0.1-1 (2011-09-29)

• First CRAN release.

• See vignette("partykit", package = "partykit") for a (somewhat rough) introduction to the package and its classes/methods.