-
Notifications
You must be signed in to change notification settings - Fork 0
/
NEWS.Rd
274 lines (237 loc) · 8.94 KB
/
NEWS.Rd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
\name{NEWS}
\title{NEWS file for the mlt package}
\section{Changes in Version 1.2-3 (2021-02-22)}{
\subsection{New features}{
\itemize{
\item{Improve documentation.}
}
}
}
\section{Changes in Version 1.2-2 (2021-02-12)}{
\subsection{Bugfixes}{
\itemize{
\item{Sampling from unconditional models did not pay attention to
number of observations.}
}
}
}
\section{Changes in Version 1.2-1 (2021-02-03)}{
\subsection{New features}{
\itemize{
\item{Quantiles and thus simulations are now computationally more exact and more robust.
The unnecessary \code{interpolate} argument to \code{predict} and
\code{simulate} is now ignored.}
\item{Adjust contrasts a fixed parameter contributes to.}
\item{Return numerically determined Hessians upon request.}
\item{Implement frailty error distributions, experimentally and internal only.}
\item{Implement cure mixture models, experimentally and internal only.}
\item{Improve computations of log-probabilities.}
}
}
\subsection{Bugfixes}{
\itemize{
\item{Discrete hazard functions were incorrect.}
}
}
}
\section{Changes in Version 1.2-0 (2020-05-11)}{
\itemize{
\item{Add exponential distribution (for Aalen additive hazards models).}
\item{Pay attention to model class when computing cumulative hazards.}
\item{Add log-cumulative hazards, log-odds, and odds for predictions and plots.}
}
}
\section{Changes in Version 1.1-2 (2020-02-13)}{
\itemize{
\item{Allow permutations of single variables.}
\item{Update citation info.}
}
}
\section{Changes in Version 1.1-1 (2020-01-09)}{
\itemize{
\item{Try harder to invert Hessians.}
\item{Update reference output.}
}
}
\section{Changes in Version 1.1-0 (2019-11-06)}{
\itemize{
\item{Add support for \pkg{nloptr} (still experimental and thus switched off by default).}
\item{Make sure \code{coef()} always returns named argument.}
\item{Fix problem in \code{as.Surv} reported by Balint Tamasi.}
}
}
\section{Changes in Version 1.0-7 (2019-10-07)}{
\itemize{
\item{Less paranoia in \file{bugfixes.R}.}
}
}
\section{Changes in Version 1.0-6 (2019-10-04)}{
\itemize{
\item{Return Hessian for fixed parameters if requested.}
\item{Fix subsetting problem in \code{R.numeric}.}
\item{Allow to \code{update} offsets.}
}
}
\section{Changes in Version 1.0-5 (2019-06-20)}{
\itemize{
\item{Add a \code{bread} method.}
\item{Check response variable against observations in \code{data}.}
\item{Make sure integers larger zero are handled correctly in \code{R}.}
\item{Implement \code{resid} method, ie the score wrt a constant.}
\item{Cox examples with Bernstein polynomials of log-time.}
}
}
\section{Changes in Version 1.0-4 (2018-12-03)}{
\itemize{
\item{Arguments \code{K} and \code{cheat} where
ignored by \code{confband} when \code{newdata}
had multiple rows.}
\item{Computation of starting values more robust now.}
\item{Order of fixed parameters (\code{fixed} argument to \code{mlt})
might have been wrong due to incomplete matching.}
}
}
\section{Changes in Version 1.0-3 (2018-09-17)}{
\itemize{
\item{Add \code{lty} argument to \code{plot.ctm}.}
\item{\code{update} needs free coefficients only.}
\item{Internal interface changes.}
}
}
\section{Changes in Version 1.0-2 (2018-06-22)}{
\itemize{
\item{Make sure transformation functions outside \code{bounds} are minus
or plus \code{Inf}.}
\item{Initial guestimates for ordered responses were incorrect
and may potentially have led to nonsense results.}
\item{Some smaller improvements in computation of log-likelihoods and
scores with respect to accuracy and speed.}
\item{\code{print} respects \code{options(digits)}.}
}
}
\section{Changes in Version 1.0-1 (2018-04-20)}{
\itemize{
\item{\code{estfun, parm = coef(object, fixed = TRUE))} evaluates
scores for all model parameters, including fixed ones.}
\item{\code{logLik(..., newdata, w)} ignored weights \code{w} when
\code{newdata} was given. Same problem was also fixed for
\code{estfun}.}
}
}
\section{Changes in Version 1.0-0 (2018-03-07)}{
\itemize{
\item{A paper describing version 1.0-0 of the \pkg{mlt}, \pkg{basefun}, and
\pkg{variables} packages was accepted for publication in the Journal of
Statistical Software 2018-03-05.}
\item{Documentation updates.}
}
}
\section{Changes in Version 0.2-2 (2018-02-23)}{
\itemize{
\item{Use \pkg{coneprog} for getting the starting values.}
\item{\code{logLik} and \code{estfun} accept matrices as \code{parm}
argument for the evalution of log-likelihoods and scores
with subject-specific parameters (for example in transformation
trees or forests and boosting procedures.}
}
}
\section{Changes in Version 0.2-1 (2017-12-12)}{
\itemize{
\item{\code{q} is forwarded to \code{qmlt} by \code{predict.ctm} now.}
\item{\code{p} is now \code{prob} in \code{qmlt} and thus \code{predict.ctm}.}
\item{Update citation.}
}
}
\section{Changes in Version 0.2-0 (2017-06-20)}{
\itemize{
\item{Most Likely Transformations will be published in
the Scandinavian Journal of Statistics.}
\item{Import package \pkg{alabama}.}
\item{\code{as.Surv(R(Surv(...)))} returns \code{Surv(...)},
useful for converting output by \code{simulate}
to \code{Surv} objects.}
}
}
\section{Changes in Version 0.1-3 (2017-04-04)}{
\itemize{
\item{Add \code{subset} argument to \code{update} (for faster transformation
trees and forests).}
\item{Sum over score contributions with positive weight only when
evaluating the gradient.}
\item{Having all response observations being interval-censored is
allowed again (too heavy checking was in place).}
\item{Don't try to numerically check KKT conditions automatically.}
\item{Check for unused arguments in dots where necessary.}
\item{Make sure the score doesn't get too large (avoid division by near
zero probabilities).}
\item{Improve \code{survfit} to compute non-parametric unconditional
probabilities for obtaining starting values in the presence of
censoring and truncation.}
}
}
\section{Changes in Version 0.1-2 (2017-02-14)}{
\itemize{
\item{\code{logLik} with \code{newdata} argument ignored \code{parm}
and \code{weights} arguments.}
\item{\code{estfun} now also has a \code{newdata} argument.}
\item{Correct axes labelling when plotting quantile functions.}
\item{make sure names are correct in \code{coef(model, fixed = FALSE)}.}
\item{check if any exact or interval-censored response with non-zero
weight exists before trying to fit the model.}
\item{make checks a little more robust against huge diffs under Windows.}
}
}
\section{Changes in Version 0.1-1 (2016-11-29)}{
\itemize{
\item{Fix two bugs in computation of log-likelihood for possibly
left-truncated responses such as \code{Surv(start, time, status)}.}
}
}
\section{Changes in Version 0.1-0 (2016-11-10)}{
\itemize{
\item{Add augmented lagrangian minimization (\code{auglag()} from package \pkg{alabama}).}
\item{Make optimiation procedure more general and adaptive, allow users to change
defaults and even add their own optimiser.}
}
}
\section{Changes in Version 0.0-34 (2016-09-13)}{
\itemize{
\item{fix bug when calling \code{survfit} for computing initial probabilities.}
\item{add \code{bysim} argument to \code{simulate}.}
\item{make sure \code{checkGrad} is respected by \code{update}.}
\item{\code{predict} computes \code{q} with \code{K} elements
if not given (as \code{plot} always did).}
}
}
\section{Changes in Version 0.0-33 (2016-05-25)}{
\itemize{
\item{Make sure \code{times} are ordered before calling
\code{survival::summary.survfit}.}
}
}
\section{Changes in Version 0.0-32 (2016-05-24)}{
\itemize{
\item{Introduce \code{as.mlt} generic.}
\item{Introduce a \code{coef} slot in \code{ctm} objects
and a corresponding \code{coef<-} and \code{coef}
method for setting and extracting coefficients
to and from unfitted conditional transformation models.}
\item{\code{predict}, \code{simulate} and \code{plot}
work on \code{ctm} objects (with meaningful coefficients)
now.}
}
}
\section{Changes in Version 0.0-31 (2016-03-11)}{
\itemize{
\item{Some small improvements wrt run time and memory consumption.}
\item{Use \code{theta = coef(object)} as default starting parameters in \code{update()}.}
\item{\code{logLik} has a new \code{newdata} argument.}
\item{\code{simulate} has a new \code{q} argument.}
}
}
\section{Changes in Version 0.0-30 (2016-02-19)}{
\itemize{
\item{The \pkg{mlt} package was first
published on CRAN.}
}
}