2025-01-12 04:36:52 +08:00

147 lines
5.0 KiB
R

\name{NEWS}
\title{NEWS file for the rpart package}
\section{Changes in version 4.1-14}{
\itemize{
\item Changed example data solder to solder.balance. The full
version of the data is available in the survival package.
}}
\section{Changes in version 4.1-10}{
\itemize{
\item Rpart would fail with a formula having ~. - x on the right hand
side. A simple bookkeeping error in creating an index.
\item Added a section to the vignette on user written functions,
which explains why and when one can avoid checking all 2^k
splits for a categorical predictor with k levels.
}}
\section{Changes in version 4.1-0}{
\itemize{
\item The C and R code has been reformatted for legibility.
\item The old compatibility function \code{rpconvert()} has been removed.
\item The cross-validation functions allow for user interrupt at the
end of evaluating each split.
\item Variable \code{Reliability} in data set \code{car90} is
corrected to be an ordered factor, as documented.
\item Surrogate splits are now considered only if they send two or
more cases \emph{with non-zero weight} each way. For
numeric/ordinal variables the restriction to non-zero weights is
new: for categorical variables this is a new restriction.
\item Surrogate splits which improve only by rounding error over the
default split are no longer returned. Where weights and missing
values are present, the \code{splits} component for some of these
was not returned correctly.
}
}
\section{Changes in version 4.0-3}{
\itemize{
\item A fit of class \samp{"rpart"} now contains a component for
variable \sQuote{importance}, which is reported by the
\code{summary()} method.
\item The \code{text()} method gains a \code{minlength} argument,
like the \code{labels()} method. This adds finer control: the
default remains \code{pretty = NULL}, \code{minlength = 1L}.
\item The handling of fits with zero and fractional weights has been
corrected: the results may be slightly different (or even
substantially different when the proportion of zero weights is
large).
\item Some memory leaks have been plugged.
\item There is a second vignette, \file{longintro.Rnw}, a version of
the original Mayo Tecnical Report on \pkg{rpart}.
}
}
\section{Changes in version 4.0-2}{
\itemize{
\item Added dataset \code{car90}, a corrected version of the
S-PLUS dataset \code{car.all} (used with permission).
\item This version does not use \code{paste0{}} and so works
with \R 2.14.x.
}
}
\section{Changes in version 4.0-1}{
\itemize{
\item Merged in a set of Splus code changes that had accumulated at
Mayo over the course of a decade. The primary one is a change in how
indexing is done in the underlying C code, which leads to a major
speed increase for large data sets. Essentially, for the lower
leaves all our time used to be eaten up by bookkeeping, and this was
replaced by a different approach. The primary routine also uses
\code{.Call{}} so as to be more memory efficient.
\item The other major change was an error for asymmetric loss
matrices, prompted by a user query. With L=loss asymmetric, the
altered priors were computed incorrectly -- they were using L'
instead of L. Upshot -- the tree would not not necessarily choose
optimal splits for the given loss matrix. Once chosen, splits were
evaluated correctly. The printed \dQuote{improvement} values are of
course the wrong ones as well. It is interesting that for my little
test case, with L quite asymmetric, the early splits in the tree are
unchanged -- a good split still looks good.
\item Add the \code{return.all} argument to \code{xpred.rpart()}.
\item Added a set of formal tests, i.e., cases with known answers to
which we can compare.
\item Add a \file{usercode} vignette, explaining how to add user defined
splitting functions.
\item The class method now also returns the node probability.
\item Add the \code{stagec} data set, used in some tests.
\item The \code{plot.rpart} routine needs to store a value that will
be visible to the \code{rpartco} routine at a later time. This is
now done in an environment in the namespace.
}
}
\section{Changes in version 3.1-55}{
\itemize{
\item Force use of registered symbols in R >= 2.16.0
\item Update Polish translations.
\item Work on message formats.
}
}
\section{Changes in version 3.1-54}{
\itemize{
\item Add Polish translations
}
}
\section{Changes in version 3.1-53}{
\itemize{
\item \code{rpart}, \code{rpart.matrix}: allow backticks in formulae.
\item \file{tests/backtick.R}: regession test
}
}
\section{Changes in version 3.1-52}{
\itemize{
\item \file{src/xval.c}: ensure unused code is not compiled in.
}
}
\section{Changes in version 3.1-51}{
\itemize{
\item Change description of \samp{margin} in \code{?plot.rpart}
as suggested by Bill Venables.
}
}