SCM

SCM Repository

[surveillance] Annotation of /pkg/inst/NEWS.Rd
ViewVC logotype

Annotation of /pkg/inst/NEWS.Rd

Parent Directory Parent Directory | Revision Log Revision Log


Revision 2253 - (view) (download) (as text)

1 : wastl 1792 %% Some pre-defined commands: \R, \code, \acronym, \url, \file, \pkg
2 :     %% Since R 3.2.0, additional system Rd macros are available,
3 :     %% e.g., \CRANpkg and \doi. See the definitions in the file
4 :     %% file.path(R.home("share"), "Rd", "macros", "system.Rd")
5 :    
6 :     \name{NEWS}
7 :     \title{News for Package 'surveillance'}
8 :     \encoding{latin1}
9 :    
10 :    
11 : wastl 2207 \section{Changes in surveillance version 1.16.2.9000}{
12 :    
13 : wastl 2237 \subsection{NEW FEATURES}{
14 :     \itemize{
15 :     \item \code{W_powerlaw(..., from0 = TRUE)} enables more parsimonious
16 : wastl 2246 \code{hhh4} models in that the power-law weights are modified to
17 :     include the autoregressive (0-distance) case (see
18 :     \code{vignette("hhh4_spacetime")}). The unstructured distance
19 :     weights \code{W_np()} gained \code{from0} support as well.
20 : wastl 2237 }
21 :     }
22 :    
23 : wastl 2238 \subsection{MINOR CHANGES}{
24 :     \itemize{
25 :     \item \code{W_np()}'s argument \code{to0} has been renamed to
26 :     \code{truncate}. The old name still works but is deprecated.
27 :     }
28 :     }
29 :    
30 : wastl 2207 \subsection{BUG FIXES}{
31 :     \itemize{
32 : wastl 2253 \item For \code{farringtonFlexible()} with multivariate time series,
33 :     only the last unit had stored the additional control items
34 :     (exceedence scores, p-values, ...), all others were 0.
35 :     [reported by Johannes Bracher]
36 :    
37 : wastl 2207 \item Using negative indices to exclude time points from an
38 :     \code{"sts"} object (e.g., \code{x[-1,]}) is now supported and
39 :     equivalent to the corresponding subset expression of retained
40 :     indexes (\code{x[2:nrow(x),]}) in resetting the \code{start} and
41 :     \code{epoch} slots. [reported by Johannes Bracher]
42 : wastl 2213
43 : wastl 2215 \item The supplementary p-values returned by \code{farringtonFlexible()}
44 :     were wrong for the default approach, where \code{thresholdMethod="delta"}
45 : wastl 2213 (i.e., the original Farrington method) and a power transformation
46 :     was applied to the data (\code{powertrans != "none"}).
47 :     [reported by Lore Merdrignac]
48 : wastl 2217
49 :     \item \code{epidataCS2sts()} no longer transfers the
50 :     \code{stgrid$BLOCK} indices to the \code{epoch} slot of the
51 :     resulting \code{"sts"} object (to avoid \code{epoch[1] != 1}
52 :     scenarios).
53 : wastl 2207 }
54 :     }
55 :    
56 :     }
57 :    
58 :    
59 : wastl 2198 \section{Changes in surveillance version 1.16.2 (2018-07-24)}{
60 : wastl 2172
61 : wastl 2178 \subsection{MINOR CHANGES}{
62 :     \itemize{
63 : wastl 2197 \item \code{autoplot.sts()} gained a \code{width} argument to adjust
64 :     the bar width, which now defaults to 7 for weekly time series
65 :     (previously was 90\% of that so there were gaps between the bars).
66 :    
67 : wastl 2178 \item \code{"epidataCS"} generation now (again) employs
68 :     \CRANpkg{spatstat}'s \code{bdist.points()}, which has been
69 :     accelerated in version 1.56-0. If you use the
70 :     \code{twinstim()}-related modelling part of \pkg{surveillance}, you
71 :     are thus advised to update your \pkg{spatstat} installation.
72 : wastl 2191
73 : wastl 2192 \item The \code{boda()} examples in
74 :     \code{vignette("monitoringCounts")} have been updated to also work
75 :     with recent versions of \pkg{INLA}.
76 : wastl 2178 }
77 :     }
78 :    
79 : wastl 2172 \subsection{BUG FIXES}{
80 :     \itemize{
81 : wastl 2192 \item Offsets in \code{hhh4}'s epidemic components were ignored by
82 : wastl 2173 \code{simulate.hhh4()} [spotted by Johannes Bracher] as well as
83 :     in dominant eigenvalues (\dQuote{maxEV}).
84 : wastl 2198
85 :     \item The color key in \code{fanplot()} is no longer distorted by
86 :     \code{log="y"}.
87 : wastl 2172 }
88 :     }
89 :    
90 :     }
91 :    
92 :    
93 : wastl 2164 \section{Changes in surveillance version 1.16.1 (2018-05-28)}{
94 : wastl 2093
95 :     \subsection{BUG FIXES}{
96 :     \itemize{
97 :     \item \code{autoplot.sts()} now sets the calling environment as
98 :     the \code{plot_env} of the result.
99 : wastl 2102
100 : wastl 2113 \item Several \code{twinstim}-related functions finally allow for
101 :     prehistory events (long supported by \code{twinstim()} itself):
102 :     \code{as.epidataCS()}, \code{glm_epidataCS()},
103 :     \code{as.epidata.epidataCS()}.
104 : wastl 2108
105 :     \item The \code{summary()} for SI[R]S-type \code{"epidata"} failed
106 :     if there were initially infectious individuals.
107 : wastl 2093 }
108 :     }
109 : wastl 2124
110 :     \subsection{DEPRECATED AND DEFUNCT}{
111 :     \itemize{
112 : wastl 2126 \item Several ancient functions have been deprecated and may be
113 :     removed in future versions of \pkg{surveillance}: \code{qlomax()},
114 : wastl 2131 \code{readData()}, \code{toFileDisProg()}, \code{correct53to52()},
115 :     \code{enlargeData()}, \code{compMatrix.writeTable()},
116 :     \code{test()}, \code{testSim()}, \code{makePlot()}.
117 : wastl 2124 }
118 : wastl 2093 }
119 :    
120 : wastl 2127 }
121 : wastl 2093
122 : wastl 2127
123 : wastl 2075 \section{Changes in surveillance version 1.16.0 (2018-01-24)}{
124 : wastl 2001
125 : wastl 2009 \subsection{NEW FEATURES}{
126 :     \itemize{
127 :     \item The \code{as.data.frame()} method for \code{"sts"} objects
128 :     gained a \code{tidy} argument, which enables conversion to the
129 : wastl 2010 long data format and is also available as function \code{tidy.sts()}.
130 : wastl 2042
131 : wastl 2059 \item A \CRANpkg{ggplot2} variant of \code{stsplot_time()} is now
132 :     available via \code{autoplot.sts()}.
133 :    
134 : wastl 2042 \item \code{as.epidata.data.frame()} gained an argument
135 :     \code{max.time} to specify the end of the observation period (which
136 :     by default coincides with the last observed event).
137 : wastl 2053
138 :     \item The now exported function \code{fanplot()} wraps
139 : wastl 2057 \CRANpkg{fanplot}\code{::fan()}. It is used by
140 :     \code{plot.oneStepAhead()} and \code{plot.hhh4sims()}, which now
141 :     have an option to add the point forecasts to the fan as well.
142 : wastl 2070
143 :     \item \code{plotHHH4_fitted()} (and \code{plotHHH4_fitted1()})
144 :     gained an option \code{total} to sum the fitted components over all
145 :     units.
146 : wastl 2009 }
147 :     }
148 :    
149 : wastl 2075 \subsection{SIGNIFICANT CHANGES}{
150 : wastl 2035 \itemize{
151 :     \item Package \CRANpkg{polyCub} is no longer automatically attached
152 : wastl 2075 (only imported).
153 : wastl 2049
154 : wastl 2057 \item \code{scores.oneStepAhead()} no longer reverses the ordering
155 :     of the time points by default, as announced in 1.15.0.
156 :     }
157 :     }
158 :    
159 :     \subsection{MINOR CHANGES}{
160 :     \itemize{
161 : wastl 2049 \item Some code in \code{vignette("monitoringCounts")} has been
162 :     adjusted to work with the new version of \CRANpkg{MGLM} (0.0.9).
163 : wastl 2067
164 :     \item Added a \code{[}-method for the \code{"hhh4sims"} class to
165 :     retain the attributes when subsetting simulations.
166 : wastl 2035 }
167 :     }
168 :    
169 : wastl 2001 \subsection{BUG FIXES}{
170 :     \itemize{
171 :     \item \code{aggregate(stsObj, by = "unit")} no longer results in
172 :     empty colnames (set to \code{"overall"}).
173 : wastl 2004 The obsolete map is dropped.
174 : wastl 2019
175 :     \item The \code{subset} argument of \code{twinSIR()} was partially
176 :     ignored:
177 :     \itemize{
178 :     \item If \code{nIntervals = 1}, the model \code{summary()}
179 :     reported the total number of events.
180 :     \item Automatic \code{knots}, model \code{residuals()}, as well as
181 :     the rug in \code{intensityplot()} were computed from the whole set
182 :     of event times.
183 :     }
184 : wastl 2036
185 :     \item The \code{as.epidata.data.frame()} converter did not actually
186 :     allow for latent periods (via \code{tE.col}). This is now possible
187 :     but considered experimental (methods for \code{"epidata"} currently
188 :     ignore latent periods).
189 : wastl 2048
190 :     \item The \code{all.equal()} methods for \code{"hhh4"} and
191 :     \code{"twinstim"} objects now first check for the correct classes.
192 : wastl 2001 }
193 :     }
194 :    
195 :     }
196 :    
197 :    
198 : wastl 1995 \section{Changes in surveillance version 1.15.0 (2017-10-06)}{
199 : wastl 1925
200 : wastl 1936 \subsection{NEW FEATURES}{
201 :     \itemize{
202 :     \item \code{siaf.gaussian()} now also employs a \code{polyCub.iso()}
203 :     integration routine by default (similar to the powerlaw-type
204 :     kernels), instead of adaptive midpoint cubature.
205 :     This increases precision and considerably accelerates estimation of
206 :     \code{twinstim()} models with a Gaussian spatial interaction
207 : wastl 1947 function. Models fitted with the new default
208 :     (\code{F.adaptive=FALSE, F.method="iso"})
209 :     will likely differ from previous fits (\code{F.adaptive=TRUE}),
210 :     and the numerical difference depends on
211 : wastl 1936 the adaptive bandwidth used before (the default \code{adapt=0.1}
212 :     yielded a rather rough approximation of the integral).
213 : wastl 1958
214 : wastl 1959 \item Added \code{quantile()}, \code{confint()}, and \code{plot()}
215 :     methods for \code{"oneStepAhead"} predictions.
216 : wastl 1995
217 :     \item Exported the function \code{simEndemicEvents()} to simulate a
218 :     spatio-temporal point pattern from an endemic-only
219 :     \code{"twinstim"}; faster than via the general
220 :     \code{simulate.twinstim()} method.
221 : wastl 1936 }
222 :     }
223 :    
224 : wastl 1930 \subsection{MINOR CHANGES}{
225 :     \itemize{
226 :     \item \code{twinstim(..., siaf = siaf.gaussian())}
227 :     uses a larger default initial value for the kernel's standard
228 : wastl 1938 deviation (based on the size of the observation region).
229 : wastl 1937
230 :     \item Non-default parametrizations of \code{siaf.gaussian()} are
231 :     deprecated, i.e., always use \code{logsd=TRUE} and \code{density=FALSE}.
232 : wastl 1948
233 :     \item \code{twinstim()} uses a smaller default initial value for the
234 :     epidemic intercept, which usually allows for faster convergence.
235 : wastl 1957
236 :     \item \code{update.hhh4()} now allows \code{subset.upper} values
237 :     beyond the originally fitted time range (but still within the time
238 :     range of the underlying \code{"sts"} object).
239 : wastl 1961
240 :     \item \code{scores.oneStepAhead()} by default reverses the ordering
241 :     of the time points. This awkward behaviour will change in the next
242 :     version, so the method now warns if the default \code{reverse=TRUE}
243 :     is used without explicit specification.
244 : wastl 1995
245 :     \item Minor improvements in the documentation and some vignettes:
246 :     corrected typos, simplified example code, documented some methods.
247 : wastl 1930 }
248 :     }
249 :    
250 : wastl 1925 \subsection{BUG FIXES}{
251 :     \itemize{
252 :     \item The C-routines introduced in version 1.14.0 used \code{==}
253 :     comparisons on parameter values to choose among case-specific
254 :     formulae (e.g., for \eqn{d==2} in \code{siaf.powerlaw()}).
255 : wastl 1936 We now employ an absolute tolerance of 1e-7 (which should fix
256 : wastl 1925 the failing tests on Solaris).
257 : wastl 1995
258 :     \item Interaction functions for \code{twinstim()}, such as
259 :     \code{siaf.powerlaw()} or \code{tiaf.exponential()}, no longer live in
260 :     the global environment as this risks using masked base functions.
261 : wastl 1925 }
262 :     }
263 :    
264 : wastl 1926 }
265 : wastl 1925
266 : wastl 1926
267 : wastl 1915 \section{Changes in surveillance version 1.14.0 (2017-06-29)}{
268 : wastl 1863
269 :     \subsection{DOCUMENTATION}{
270 :     \itemize{
271 : wastl 1915 \item The replication code from Meyer et al. (2017, JSS)
272 :     is now included as \code{demo("v77i11")}.
273 :     It exemplifies the spatio-temporal endemic-epidemic modelling
274 :     frameworks \code{twinstim}, \code{twinSIR}, and \code{hhh4}
275 :     (see also the corresponding vignettes).
276 : wastl 1863 }
277 :     }
278 :    
279 :     \subsection{NEW FEATURES}{
280 :     \itemize{
281 :     \item Pure C-implementations of integration routines for spatial
282 :     interaction functions considerably accelerate the estimation of
283 :     \code{twinstim()} models containing
284 : wastl 1881 \code{siaf.powerlaw()}, \code{siaf.powerlawL()}, or \code{siaf.student()}.
285 : wastl 1878
286 :     \item The color palette generating function used by \code{sts}
287 :     plots, \code{hcl.colors}, is now exported.
288 :    
289 :     \item The utility function \code{clapply} (\emph{c}onditional
290 :     \code{lapply}) is now exported.
291 :    
292 :     \item Some utility functions for \code{hhh4} fits are now exported
293 :     (\code{update.hhh4}, \code{getNEweights}, \code{coefW}),
294 :     as well as several internal functions for use by \code{hhh4} add-on
295 :     packages (\code{meanHHH}, \code{sizeHHH}, \code{decompose.hhh4}).
296 : wastl 1898
297 :     \item The \code{"fan"}-type plot function for \code{"hhh4sims"}
298 :     gained a \code{key.args} argument for an automatic color key.
299 : wastl 1912
300 :     \item New auxiliary function \code{makeControl()}, which may be
301 :     used to specify a \code{hhh4()} model.
302 : wastl 1863 }
303 :     }
304 :    
305 : wastl 1885 \subsection{MINOR CHANGES}{
306 :     \itemize{
307 :     \item \code{twinstim()} now throws an informative error message when
308 :     trying to fit a purely epidemic model to data containing endemic
309 :     events (i.e., events without ancestors). The \code{help("twinstim")}
310 :     exemplifies such a model.
311 :     }
312 : wastl 1863 }
313 :    
314 : wastl 1889 \subsection{BUG FIXES}{
315 :     \itemize{
316 :     \item \code{siaf.powerlaw()$deriv} returned \code{NaN} for the
317 :     partial derivative wrt the decay parameter \eqn{d}, if \eqn{d}
318 :     was large enough for \eqn{f} to be numerically equal to 0.
319 :     It will now return 0 in this case.
320 : wastl 1901
321 :     \item \code{twinstim()} could fail (with an error from
322 :     \code{duplicated.default}) if the fitted time range was
323 :     substantially reduced via the \code{T} argument.
324 : wastl 1902
325 :     \item The \code{"simEpidataCSlist"} generated by
326 :     \code{simulate.twinstim(..., simplify = TRUE)} was missing the
327 :     elements \code{bbox} and \code{control.siaf}.
328 : wastl 1889 }
329 : wastl 1885 }
330 : wastl 1863
331 : wastl 1889 }
332 : wastl 1885
333 : wastl 1889
334 : wastl 1852 \section{Changes in surveillance version 1.13.1 (2017-04-28)}{
335 : wastl 1827
336 : wastl 1852 \subsection{DOCUMENTATION}{
337 :     \itemize{
338 :     \item The paper on \dQuote{Spatio-Temporal Analysis of Epidemic
339 :     Phenomena Using the \R Package \pkg{surveillance}} (by Sebastian
340 :     Meyer, Leonhard Held, and Michael \enc{Höhle}{Hoehle}) will appear
341 :     in the upcoming volume of the \emph{Journal of Statistical Software}.
342 :     The main sections 3 to 5 of the paper are contained in the package
343 :     as \code{vignette("twinstim")}, \code{vignette("twinSIR")}, and
344 :     \code{vignette("hhh4_spacetime")}, respectively.
345 :     }
346 :     }
347 :    
348 : wastl 1827 \subsection{NEW FEATURES}{
349 :     \itemize{
350 : wastl 1829 \item The \code{calibrationTest()} and \code{pit()} methods for
351 :     \code{"oneStepAhead"} forecasts gained an argument \code{units}
352 :     to allow for unit-specific assessments.
353 : wastl 1836
354 :     \item A default \code{scores}-method is now available to compute a
355 :     set of proper scoring rules for Poisson or NegBin predictions.
356 : wastl 1842
357 :     \item New plot \code{type = "fan"} for simulations from
358 :     \code{"hhh4"} models to produce a fan chart using the
359 :     \CRANpkg{fanplot} package.
360 : wastl 1827 }
361 :     }
362 :    
363 :     \subsection{MINOR CHANGES}{
364 :     \itemize{
365 : wastl 1834 \item \code{scores.hhh4()} sets rownames for consistency with
366 :     \code{scores.oneStepAhead()}.
367 : wastl 1827 }
368 :     }
369 :    
370 :     \subsection{BUG FIXES}{
371 :     \itemize{
372 : wastl 1843 \item The \code{"Lambda.const"} matrix returned by
373 :     \code{getMaxEV_season()} was wrong for models with asymmetric
374 :     neighbourhood weights. [spotted by Johannes Bracher]\cr
375 :     Dominant eigenvalues (\code{"maxEV"}) were not affected by this bug.
376 : wastl 1827 }
377 :     }
378 :    
379 :     }
380 :    
381 :    
382 : wastl 1821 \section{Changes in surveillance version 1.13.0 (2016-12-20)}{
383 : wastl 1792
384 :     \subsection{NEW FEATURES}{
385 :     \itemize{
386 : wastl 1793 \item \code{earsC} now has two new arguments thanks to Howard
387 :     Burkom: the number of past time units to be used in calculation is
388 :     now not always 7, it can be chosen in the \code{baseline} parameter.
389 :     Furthermore, the \code{minSigma} parameter allows to get a threshold
390 :     in the case of sparse data. When one doesn't give any value for those
391 :     two parameters, the algorithm works like it used to.
392 :    
393 :     \item \code{animate.sts()} gained support for date labels in the
394 :     bottom \code{timeplot}.
395 : wastl 1802
396 :     \item \code{stsplot_space()} and \code{animate.sts()} can now
397 :     generate incidence maps based on the population information
398 :     stored in the supplied \code{"sts"} object.
399 :     Furthermore, \code{animate.sts()} now supports time-varying
400 :     population numbers.
401 : wastl 1792 }
402 :     }
403 :    
404 :     \subsection{MINOR CHANGES}{
405 :     \itemize{
406 :     \item \code{hhh4()} guards against the misuse of
407 :     \code{family = factor("Poisson")} for univariate time series.
408 :     Previously, this resulted in a negative binomial model by
409 :     definition, but is now interpreted as \code{family = "Poisson"}
410 :     (with a warning).
411 :     }
412 :     }
413 :    
414 :     \subsection{BUG FIXES}{
415 :     \itemize{
416 :     \item \code{animate.sts()} now supports objects with missing values
417 : wastl 1800 (with a warning). Furthermore, the automatic color breaks have been
418 :     improved for incidence maps, also in \code{stsplot_space()}.
419 : wastl 1792
420 :     \item The \code{as.data.frame}-method for the \code{"sts"} class,
421 :     applied to classical time-index-based \code{"sts"} objects
422 :     (\code{epochAsDate=FALSE}), ignored a \code{start} epoch different
423 :     from 1 when computing the \code{epochInPeriod} indexes.
424 :     Furthermore, the returned \code{epochInPeriod} now is a fraction of
425 :     \code{freq}, for consistency with the result for objects with
426 :     \code{epochAsDate=TRUE}.
427 : wastl 1807
428 :     \item \code{simulate.hhh4()} did not handle shared overdispersion
429 :     parameters correctly. The different parameters were simply recycled
430 :     to the number of units, ignoring the factor specification from the
431 :     model's \code{family}. [spotted by Johannes Bracher]
432 : wastl 1810
433 :     \item Simulations from \emph{endemic-only} \code{"hhh4"} models
434 :     with unit-specific overdispersion parameters used wrong
435 :     variances. [spotted by Johannes Bracher]
436 : wastl 1813
437 :     \item \code{oneStepAhead()} predictions of \code{type}
438 :     \code{"rolling"} (or \code{"first"}) were incorrect for time points
439 :     \code{tp} (\code{tp[1]}) beyond the originally fitted time range
440 :     (in that they were based on the original time range only).
441 :     This usage of \code{oneStepAhead()} was never really supported and
442 :     is now catched when checking the \code{tp} argument.
443 : wastl 1817
444 :     \item \code{plot.hhh4simslist()} ignored its \code{par.settings}
445 :     argument if \code{groups=NULL} (default).
446 : wastl 1792 }
447 :     }
448 :    
449 :     }
450 :    
451 :    
452 :     \section{Changes in surveillance version 1.12.2 (2016-11-14)}{
453 :    
454 :     \subsection{NEW FEATURES}{
455 :     \itemize{
456 :     \item The internal auxiliary function, which determines the sets of
457 :     potential source events in \code{"epidataCS"} has been implemented
458 :     in \samp{C++}, which accelerates \code{as.epidataCS()},
459 :     \code{permute.epidataCS()}, and therefore \code{epitest()}.
460 :     This is only really relevant for \code{"epidataCS"} with a large
461 :     number of events (>1000, say).
462 :    
463 :     \item Negative-binomial \code{hhh4()} models may not converge for
464 :     non-overdispersed data (try, e.g.,
465 :     \code{set.seed(1); hhh4(sts(rpois(104, 10)), list(family="NegBin1"))}).
466 :     The resulting non-convergence warning message now mentions low
467 :     overdispersion if this is detected. [suggested by Johannes Bracher]
468 :    
469 :     \item An additional \code{type="delay"} option was added to the
470 :     \code{plot} method of \code{stsNC} objects. Furthermore, an
471 :     \code{animate_nowcasts} function allows one to animate a sequence of
472 :     nowcasts.
473 :    
474 :     }
475 :     }
476 :    
477 :     \subsection{MINOR CHANGES}{
478 :     \itemize{
479 :     \item In the \code{animate}-method for \code{"sts"} objects,
480 :     the default top padding of \pkg{lattice} plots is now disabled for the
481 :     bottom \code{timeplot} to reduce the space between the panels.
482 :     Furthermore, the new option \code{fill} can be used to make the
483 :     panel of the \code{timeplot} as large as possible.
484 :     }
485 :     }
486 :    
487 :     \subsection{BUG FIXES}{
488 :     \itemize{
489 :     \item \code{bodaDelay()}: fixed spurious warnings from \code{rnbinom()}.
490 :     \item \code{vignette("monitoringCounts")}: fixed \code{boda}-related
491 :     code and cache to obtain same results as in corresponding JSS paper.
492 :     }
493 :     }
494 :    
495 :     }
496 :    
497 :    
498 :    
499 :     \section{Changes in surveillance version 1.12.1 (2016-05-18)}{
500 :    
501 :     \subsection{DOCUMENTATION}{
502 :     \itemize{
503 :     \item The new \code{vignette("monitoringCounts")} illustrates the
504 :     monitoring of count time series in \R with a particular focus on
505 :     aberration detection in public health surveillance.
506 :     This vignette corresponds to a recently accepted manuscript
507 :     for the \emph{Journal of Statistical Software}
508 :     (Salmon, Schumacher, and \enc{Höhle}{Hoehle}, 2016).
509 :     }
510 :     }
511 :    
512 :     \subsection{MINOR CHANGES}{
513 :     \itemize{
514 :     \item Non-convergent \code{hhh4()} fits now obey the structure of
515 :     standard \code{"hhh4"} objects. In particular, such fits now also
516 :     contain the \code{control} and \code{stsObj} elements, allowing
517 :     for model \code{update()}s of non-convergent fits.
518 :    
519 :     \item \code{knox()} warns about symmetric input matrices.
520 :     }
521 :     }
522 :    
523 :     \subsection{BUG FIXES}{
524 :     \itemize{
525 :     \item The code of \code{boda()} (with \code{samplingMethod="joint"})
526 :     and \code{bodaDelay()} (with \code{inferenceMethod="INLA"})
527 :     has been adjusted to a change of arguments of \pkg{INLA}'s
528 :     \code{inla.posterior.sample} function. Accordingly, the minimum
529 :     \pkg{INLA} version required to run \code{boda()} and
530 :     \code{bodaDelay()} is 0.0-1458166556.
531 :    
532 :     \item The functions returned by \code{W_powerlaw()} now have the
533 :     package namespace as their environment to support situations where
534 :     the package is not attached.
535 :    
536 :     \item Attaching package \CRANpkg{nlme} after \pkg{surveillance} no
537 :     longer masks \code{"hhh4"}'s \code{ranef}-method. (We now import the
538 :     \code{fixef} and \code{ranef} generics from \pkg{nlme}.)
539 :     }
540 :     }
541 :    
542 :     }
543 :    
544 :    
545 :    
546 :     \section{Changes in surveillance version 1.12.0 (2016-04-02)}{
547 :    
548 :     \subsection{DOCUMENTATION}{
549 :     \itemize{
550 :     \item Several new vignettes illustrate \emph{endemic-epidemic}
551 :     modeling frameworks for spatio-temporal surveillance data:
552 :     \describe{
553 :     \item{\code{vignette("twinstim")}}{describes a spatio-temporal
554 :     point process regression model.}
555 :     \item{\code{vignette("twinSIR")}}{describes a multivariate
556 :     temporal point process regression model.}
557 :     \item{\code{vignette("hhh4_spacetime")}}{describes an areal
558 :     time-series model for infectious disease counts.}
559 :     }
560 :     These vignettes are based on a recently accepted manuscript
561 :     for the \emph{Journal of Statistical Software}
562 :     (Meyer, Held, and \enc{Höhle}{Hoehle}, 2016).
563 :     \item Improved the documentation on various help pages.
564 :     \item The \code{hhh4()}-based analysis of \code{data("fluBYBW")}
565 :     has been moved to a separate demo script \file{fluBYBW.R}. Due to
566 :     the abundance of models and the relatively long runtime, we
567 :     recommend to open the script in an editor rather than running
568 :     all the code at once using \code{demo("fluBYBW")}.
569 :     }
570 :     }
571 :    
572 :     \subsection{NEW FEATURES}{
573 :     \itemize{
574 :     \item Overhaul of the \code{"sts"} implementation. This mostly
575 :     affects package-internal code, which is simpler, cleaner and better
576 :     tested now, but requires \R >= 3.2.0 (due to \code{callNextMethod()}
577 :     bugs in older versions of \R).
578 :     Beyond that, the user-level constructor function \code{sts()} now
579 :     has explicit arguments for clarity and convenience.
580 :     For instance, its first argument sets the \code{observed} slot and
581 :     no longer needs to be named, i.e.,
582 :     \code{sts(mycounts, start=c(2016,3), frequency=12)}
583 :     works just like for the classical \code{ts()} function.
584 :    
585 :     \item \code{stsplot_time(..., as.one=TRUE)} is now implemented
586 :     (yielding a simple \code{matplot} of multiple time series).
587 :     }
588 :     }
589 :    
590 :     \subsection{MINOR CHANGES}{
591 :     \itemize{
592 :     \item \code{plotHHH4_season()} now by default draws a horizontal
593 :     reference line at unity if the multiplicative effect of component
594 :     seasonality is shown (i.e., if \code{intercept=FALSE}).
595 :    
596 :     \item Since \pkg{surveillance} 1.8-0, \code{hhh4()} results are of
597 :     class \code{"hhh4"} instead of \code{"ah4"} (renamed).
598 :     Legacy methods for the old class name \code{"ah4"} have been removed.
599 :    
600 :     \item The internal model preparation in \code{twinstim()} is more
601 :     efficient (the distance matrix of the events is only computed if
602 :     event sources actually need to be updated).
603 :     }
604 :     }
605 :    
606 :     \subsection{BUG FIXES}{
607 :     \itemize{
608 :     \item \code{stsplot_spacetime()} now recognizes its \code{opts.col}
609 :     argument.
610 :    
611 :     \item Conversion from \code{"ts"} to \code{"sts"} using
612 :     \code{as(ts, "sts")} could set a wrong start time. For instance,
613 :     \code{as(ts(1:10, start=c(1959,2), frequency=4), "sts")@start} was
614 :     \code{c(1959,1)}.
615 :    
616 :     \item \code{algo.twins()} now also accepts \code{"sts"} input and
617 :     the automatic legend in the first plot of \code{plot.atwins()} works
618 :     again.
619 :    
620 :     \item The experimental \code{profile}-method for \code{"twinstim"}
621 :     objects did not work if embedded \code{twinstim()} fits issued warnings.
622 :     }
623 :     }
624 :    
625 :     }
626 :    
627 :    
628 :    
629 :     \section{Changes in surveillance version 1.11.0 (2016-02-08)}{
630 :    
631 :     \subsection{NEW FEATURES}{
632 :     \itemize{
633 :     \item \code{update.epidata()} can now handle a distance matrix
634 : wastl 1807 \code{D} in the form of a classed \code{"Matrix"}.
635 :     [suggested by George Wood]
636 : wastl 1792
637 :     \item \code{glrnb()} can now handle \code{ret="cases"} for the
638 :     generalized likelihood ratio detector based on the negative binomial
639 :     distribution. It's based on a brute-force search and hence might be
640 :     slow in some situations.
641 :    
642 :     \item \code{boda()} and \code{bodaDelay()} now support an
643 :     alternative method (\code{quantileMethod="MM"}) to compute quantiles
644 :     based on the posterior distribution. The new method samples
645 :     parameters from the posterior distribution and then computes the
646 :     quantile of the mixture distribution using bisectionning, which is
647 :     faster and yields similar results compared to the original method
648 :     (\code{quantileMethod="MC"}, still the default).
649 :     }
650 :     }
651 :    
652 :     \subsection{MINOR CHANGES}{
653 :     \itemize{
654 :     \item Revised \code{vignette("hhh4")}, updated the package
655 :     description as well as some references in the documentation.
656 :     Also updated (the cache of) the slightly outdated
657 :     \code{vignette("surveillance")} to account for the corrected version
658 :     of \code{algo.bayes()} implemented since \pkg{surveillance} 1.10-0.
659 :     }
660 :     }
661 :    
662 :     \subsection{BUG FIXES}{
663 :     \itemize{
664 :     \item Fixed bug in \code{categoricalCUSUM()}, which ignored alarms
665 :     generated for the last time point in \code{range}. Furthermore, the
666 :     exact computation in case of returns of the type \code{"value"} for
667 :     the binomial are now checked through an attribute.
668 :    
669 :     \item Fixed bug in the \code{estimateGLRNbHook} function of
670 :     \code{algo.glrnb}, which ignored potential fixed \code{alpha}
671 :     values. If \code{alpha} is fixed this is now taken into consideration
672 :     while fitting the negative binomial function. See revised help files
673 :     for the details.
674 :    
675 :     \item Made a hot-fix such that the \code{algo.quality} function now
676 :     also works for \code{sts} objects and if the \code{state} or
677 :     \code{alarm} slots consists of TRUE/FALSE instead of 0/1.
678 :    
679 :     \item \code{intensity.twinstim()} did not work for non-endemic models.
680 :    
681 :     \item A parallelized \code{epitest()} could fail with a strange
682 :     error message if some replications were left unassigned.
683 :     This seems to happen if forking is used (\code{mclapply}) with
684 :     insufficient memory. Incomplete replications are now ignored with a
685 :     warning.
686 :     }
687 :     }
688 :    
689 :     }
690 :    
691 :    
692 :    
693 :     \section{Changes in surveillance version 1.10-0 (2015-11-04)}{
694 :    
695 :     \subsection{NEW FEATURES}{
696 :     \itemize{
697 :     \item Calibration tests for count data (Wei and Held, 2014, Test)
698 :     are now implemented and available as \code{calibrationTest()}.
699 :     In addition to a default method taking pure counts and predictive
700 :     means and dispersion parameters, there are convenient methods for
701 :     \code{"hhh4"} and \code{"oneStepAhead"} objects.
702 :    
703 :     \item Shared overdispersion across units in negative binomial
704 :     \code{hhh4()} time series models (by specifying a factor variable
705 :     as the \code{family} argument).
706 :    
707 :     \item \code{scores()} and \code{pit()} are now generic and have convenient
708 :     methods for \code{"oneStepAhead"} predictions and \code{"hhh4"} fits.
709 :    
710 :     \item The initial values used for model updates during the
711 :     \code{oneStepAhead()} procedure can now be specified directly
712 :     through the \code{which.start} argument (as an alternative to the
713 :     previous options \code{"current"} and \code{"final"}).
714 :    
715 :     \item \code{plotHHH4_fitted()} (and \code{plotHHH4_fitted1()})
716 :     gained an option \code{decompose} to plot the contributions from
717 :     each single unit (and the endemic part) instead of the default
718 :     endemic + AR + neighbours decomposition.
719 :     Furthermore, a formatted time axis similar to \code{stsplot_time1()}
720 :     can now be enabled via the new argument \code{xaxis}.
721 :    
722 :     \item The new \code{plot} \code{type} \code{"maps"} for \code{"hhh4"}
723 :     fits shows maps of the fitted mean components averaged over time.
724 :    
725 :     \item New \code{plot}-method for simulations from \code{"hhh4"}
726 :     models (using \code{simulate.hhh4(..., simplify = TRUE)}, which now
727 :     has a dedicated class: \code{"hhh4sims"}) to show the final
728 :     size distribution or the simulated time series
729 :     (possibly stratified by groups of units).
730 :     There is also a new \code{scores}-method to compute proper scoring
731 :     rules based on such simulations.
732 :    
733 :     \item The argument \code{idx2Exp} of \code{coef.hhh4()} may now be
734 :     conveniently set to \code{TRUE} to exp-transform all coefficients.
735 :    
736 :     \item Added a \code{coeflist()}-method for \code{"hhh4"} fits.
737 :    
738 :     \item The generator function \code{sts()} can now be used to
739 :     initialize objects of class \code{"sts"} (instead of writing
740 :     \code{new("sts", ...)}).
741 :    
742 :     \item Additional arguments of \code{layout.scalebar()} now allow to
743 :     change the style of the labels.
744 :    
745 :     \item A pre-computed distance matrix \code{D} can now be used as
746 :     input for the \code{as.epidata()} converter -- offering an alternative
747 :     to the default Euclidean distance based on the individuals coordinates.
748 :     (Request of George Wood to support \code{twinSIR} models on networks.)
749 :     }
750 :     }
751 :    
752 :     \subsection{MINOR CHANGES}{
753 :     \itemize{
754 :     \item The first argument of \code{scores()} is now called \code{x}
755 :     instead of \code{object} (for consistency with \code{calibrationTest()}).
756 :    
757 :     \item The result of \code{oneStepAhead()} now has the dedicated
758 :     class attribute \code{"oneStepAhead"} (previously was just a list).
759 :    
760 :     \item Changed interpretation of the \code{col} argument of
761 :     \code{plotHHH4_fitted()} and \code{plotHHH4_fitted1()} (moved color
762 :     of \dQuote{observed} to separate argument \code{pt.col} and reversed
763 :     remaining colors). The old \code{col} specification as a vector of
764 :     length 4 still works (catched internally) but is undocumented.
765 :    
766 :     \item The \code{epoch} slot of class \code{"sts"} is now initialized to
767 :     \code{1:nrow(observed)} by default and thus no longer needs to be
768 :     explicitly set when creating a \code{new("sts", ...)} for this
769 :     standard case.
770 :    
771 :     \item Initialization of \code{new("sts", ...)} now supports the
772 :     argument \code{frequency} (for consistency with \code{ts()}).
773 :     Note that \code{freq} still works (via partial argument matching)
774 :     and that the corresponding \code{"sts"} slot is still called \code{freq}.
775 :    
776 :     \item If \code{missing(legend.opts)} in \code{stsplot_time1()}, the
777 :     default legend will only be produced if the \code{"sts"} object
778 :     contains information on outbreaks, alarms, or upperbounds.
779 :    
780 :     \item The default \code{summary()} of a \code{"twinstim"} fit is more
781 :     concise since it no longer includes the number of log-likelihood and
782 :     score function evaluations and the elapsed time during model fitting.
783 :     Set the new \code{runtime} argument of \code{summary.twinstim()} to
784 :     \code{TRUE} to add this information to the summary as before.
785 :    
786 :     \item The \code{animate}-method for \code{"sts"} objects gained an
787 :     argument \code{draw} (to disable the default instantaneous plotting)
788 :     and now invisibly returns the sequential plot objects (of class
789 :     \code{"gtable"} or \code{"trellis"}) in a list for post-processing.
790 :    
791 :     \item The flexible time axis configurations for \code{"sts"} plots
792 :     introduced in version 1.8-0 now also work for classical \code{"sts"}
793 :     objects with integer epochs and standard frequencies
794 :     (try \code{plot(..., epochsAsDate = TRUE)}).
795 :    
796 :     \item \code{stsplot_time()} initiates \code{par} settings only
797 :     if the \code{par.list} argument is a list.
798 :    
799 :     \item The new \code{all.equal()} method for class \code{"hhh4"}
800 :     compares two fits ignoring their \code{"runtime"} and \code{"call"}
801 :     elements (at least).
802 :     }
803 :     }
804 :    
805 :     \subsection{BUG FIXES}{
806 :     \itemize{
807 :     \item Fixed a bug in \code{algo.bayes}, where an alarm was already
808 :     sounded if the current observation was equal to the quantile of the
809 :     predictive posterior. This was changed in order to get \eqn{alarm_t
810 :     = I(obs_t > quantile_t)} which is consistent with the use in
811 :     \code{boda} and \code{bodaDelay}.
812 :    
813 :     \item Fixed bug in \code{algo.outbreakP} causing a halt in the
814 :     computations of \code{value="cases"} when
815 :     \code{calc.outbreakP.statistic} returned \code{NaN}. Now, a
816 :     \code{NaN} is returned.
817 :    
818 :     \item \code{wrap.algo} argument \code{control.hook} used
819 :     \code{control} argument defined outside it's scope (and not the one
820 :     provided to the function). It is now added as additional 2nd argument
821 :     to the \code{control.hook} function.
822 :    
823 :     \item \code{stsplot_time()} did not account for the optional
824 :     \code{units} argument for multivariate \code{"sts"} objects
825 :     when choosing a suitable value for \code{par("mfrow")}.
826 :    
827 :     \item \code{hhh4()} could have used a function \code{dpois()} or
828 :     \code{dnbinom()} from the global environment instead of the
829 :     respective function from package \pkg{stats}.
830 :    
831 :     \item The default time variable \code{t} created as part of the
832 :     \code{data} argument in \code{hhh4()} was incompatible with
833 :     \code{"sts"} objects having \code{epochAsDate=TRUE}.
834 :    
835 :     \item A consistency check in \code{as.epidata.default()} failed for
836 :     SI-type data (and, more generally, for all data which ended with an
837 : wastl 1807 I-event in the last time block). [spotted by George Wood]
838 : wastl 1792 }
839 :     }
840 :    
841 :     }
842 :    
843 :    
844 :    
845 :     \section{Changes in surveillance version 1.9-1 (2015-06-12)}{
846 :    
847 :     \itemize{
848 :     \item This is a quick patch release to make the test suite run
849 :     smoothly on CRAN's Windows and Solaris Sparc systems.
850 :    
851 :     \item The new \code{hhh4()} option to scale neighbourhood weights
852 :     did not work for parametric weights with more than one parameter
853 :     if \code{normalize=FALSE}.
854 :     }
855 :    
856 :     }
857 :    
858 :    
859 :    
860 :     \section{Changes in surveillance version 1.9-0 (2015-06-09)}{
861 :    
862 :     \subsection{NEW FEATURES}{
863 :     \itemize{
864 :     \item New functions and data for Bayesian outbreak detection in the
865 :     presence of reporting delays (Salmon et al., 2015):
866 :     \code{bodaDelay()}, \code{sts_observation()}, and \code{sts_creation()}.
867 :    
868 :     \item New functions implementing tests for space-time interaction:
869 :     \itemize{
870 :     \item \code{knox()} supports both the Poisson approximation and a
871 :     Monte Carlo permutation approach to determine the p-value,
872 :     \item \code{stKtest()} wraps space-time K-function methods from
873 :     package \CRANpkg{splancs} for use with \code{"epidataCS"},
874 :     \item and \code{epitest()} for \code{twinstim} models
875 :     (makes use of the new auxiliary function \code{simpleR0()}).
876 :     }
877 :    
878 :     \item New function \code{plapply()}: a parallel and verbose version
879 :     of \code{lapply()} wrapping around both \code{mclapply()} and
880 :     \code{parLapply()} of package \pkg{parallel}.
881 :    
882 :     \item New converter \code{as.xts.sts()} to transform \code{"sts"}
883 :     objects to the quasi standard \code{"xts"} class, e.g., to make use
884 :     of package \CRANpkg{dygraphs} for interactive time series plots.
885 :    
886 :     \item New options for scaling and normalization of neighbourhood
887 :     weights in \code{hhh4()} models.
888 :    
889 :     \item New auxiliary function \code{layout.scalebar()} for use as part
890 :     of \code{sp.layout} in \code{spplot()} or in the traditional
891 :     graphics system.
892 :     }
893 :    
894 :     \subsection{New features for \code{"epidataCS"}}{
895 :     \itemize{
896 :     \item New argument \code{by} for \code{plot.epidataCS()}, which
897 :     defines a stratifying variable for the events (default is the event
898 :     type as before). It can also be set to \code{NULL} to make the plot
899 :     not distinguish between event types.
900 :    
901 :     \item The spatial plot of \code{"epidataCS"} gained the arguments
902 :     \code{tiles}, \code{pop} and \code{sp.layout}, and can now produce
903 :     an \code{spplot()} with the tile-specific population levels behind
904 :     the point pattern.
905 :    
906 :     \item New function \code{permute.epidataCS()} to randomly permute
907 :     time points or locations of the events (holding other marks fixed).
908 :     }
909 :     }
910 :    
911 :     \subsection{New features for \code{twinstim()}}{
912 :     \itemize{
913 :     \item New S3-generic \code{coeflist()} to list model coefficients by
914 :     component. It currently has a default method and one for
915 :     \code{"twinstim"} and \code{"simEpidataCS"}.
916 :    
917 :     \item New argument \code{newcoef} for \code{simulate.twinstim()} to
918 :     customize the model parameters used for the simulation.
919 :    
920 :     \item New argument \code{epilink} for \code{twinstim()}, offering
921 :     experimental support for an identity link for the epidemic
922 :     predictor. The default remains \code{epilink = "log"}.
923 :    
924 :     \item Simulation from \code{"twinstim"} models and generation of
925 :     \code{"epidataCS"} is slightly faster now (faster \pkg{spatstat}
926 :     functions are used to determine the distance of events to the border).
927 :    
928 :     \item New option \code{scaled = "standardized"} in \code{iafplot()}
929 :     to plot \eqn{f(x) / f(0)} or \eqn{g(t) / g(0)}, respectively.
930 :     }
931 :     }
932 :     }
933 :    
934 :     \subsection{MINOR CHANGES}{
935 :     \itemize{
936 :     \item Initial data processing in \code{twinstim()} is faster
937 :     since event sources are only re-determined if there is effective
938 :     need for an update (due to subsetting or a change of
939 :     \code{qmatrix}).
940 :    
941 :     \item \code{formatPval()} disables \code{scientific} notation by default.
942 :    
943 :     \item The \code{"time"} plot for \code{"epidataCS"} uses the
944 :     temporal grid points as the default histogram \code{breaks}.
945 :    
946 :     \item The special \code{fe()} function which sets up fixed effects
947 :     in \code{hhh4()} models gained an argument \code{unitSpecific} as a
948 :     convenient shortcut for \code{which = rep(TRUE, nUnits)}.
949 :    
950 :     \item The convenient \code{plot} option of \code{permutationTest()}
951 :     uses \CRANpkg{MASS}::\code{truehist()} instead of \code{hist()} and
952 :     accepts graphical parameters to customize the histogram.
953 :     }
954 :     }
955 :    
956 :     \subsection{BUG FIXES}{
957 :     \itemize{
958 :     \item The \code{bodaFit} function did not draw samples from the
959 :     joint posterior. Instead draws were from the respective posterior
960 :     marginals. A new argument \code{samplingMethod} is now introduced
961 :     defaulting to the proper 'joint'. For backwards compatibility use
962 :     the value 'marginal'.
963 :    
964 :     \item The functions \code{as.epidataCS()} and \code{simEpidataCS()}
965 :     could throw inappropriate warnings when checking polygon areas
966 :     (only if \code{W} or \code{tiles}, respectively, contained holes).
967 :    
968 :     \item Non-convergent endemic-only \code{twinstim} models
969 : wastl 1807 produced an error. [spotted by Bing Zhang]
970 : wastl 1792
971 :     \item The \code{"owin"}-method of \code{intersectPolyCircle} could
972 :     have returned a rectangle-type \code{"owin"} instead of a polygon.
973 :    
974 :     \item An error occurred in \code{twinstim()} if \code{finetune=TRUE}
975 :     or choosing \code{optim()} instead of the default \code{nlminb()}
976 :     optimizer without supplying a \code{control} list in \code{optim.args}.
977 :    
978 :     \item The \code{"time"} plot for \code{"epidataCS"} did not
979 :     necessarily use the same histogram \code{breaks} for all strata.
980 :    
981 :     \item Specifying a step function of interaction via a numeric vector
982 :     of knots did not work in \code{twinstim()}.
983 :    
984 :     \item \code{plot.hhh4()} did not support an unnamed \code{type}
985 :     argument such as \code{plot(x, "season")}.
986 :    
987 :     \item \code{simEpidataCS()} did not work if \code{t0} was in the
988 :     last block of \code{stgrid} (thus it did not work for single-cell
989 :     grids), and mislabeled the \code{start} column copied to
990 :     \code{events} if there were no covariates in \code{stgrid}.
991 :    
992 :     \item Evaluating \code{intensity.twinstim()$hFUN()} at time points
993 :     before \code{t0} was an error. The function now returns
994 :     \code{NA_real_} as for time points beyond \code{T}.
995 :    
996 :     \item Truncated, normalized power-law weights for \code{hhh4()}
997 :     models, i.e., \code{W_powerlaw(maxlag = M, normalize = TRUE)}
998 :     with \code{M < max(neighbourhood(stsObj))}, had wrong derivatives
999 :     and thus failed to converge.
1000 :    
1001 :     \item \code{update.hhh4(..., use.estimates = TRUE)} did not
1002 :     use the estimated weight function parameters as initial values for
1003 :     the new fit. It does so now iff the weight function
1004 :     \code{ne$weights} is left unchanged.
1005 :     }
1006 :     }
1007 :    
1008 :     }
1009 :    
1010 :    
1011 :    
1012 :     \section{Changes in surveillance version 1.8-3 (2015-01-05)}{
1013 :    
1014 :     \itemize{
1015 :     \item Accommodate a new note given by R-devel checks, and set the new
1016 :     INLA additional repository in the \file{DESCRIPTION} file.
1017 :    
1018 :     \item Made \code{linelist2sts()} work for quarters by adding extra
1019 :     \code{"\%q"} formatting in \code{formatDate()}.
1020 :     }
1021 :    
1022 :     }
1023 :    
1024 :    
1025 :    
1026 :     \section{Changes in surveillance version 1.8-2 (2014-12-16)}{
1027 :    
1028 :     \subsection{MINOR CHANGES related to \code{hhh4}}{
1029 :     \itemize{
1030 :     \item In the coefficient vector resulting from a \code{hhh4} fit,
1031 :     random intercepts are now named.
1032 :    
1033 :     \item Parameter \code{start} values in \code{hhh4()} are now
1034 :     matched by name but need not be complete in that case (default
1035 :     initial values are used for unspecified parameters).
1036 :    
1037 :     \item The \code{update.hhh4()}-method now by default does
1038 :     \code{use.estimates} from the previous fit. This reduces the
1039 :     number of iterations during model fitting but may lead to slightly
1040 :     different parameter estimates (within a tolerance of \code{1e-5}).
1041 :     Setting \code{use.estimates = FALSE} means to re-use the previous
1042 :     start specification.
1043 :     }
1044 :     }
1045 :    
1046 :     \subsection{MINOR CHANGES related to the \code{"sts"}-class}{
1047 :     \itemize{
1048 :     \item For univariate \code{"sts"} objects, the (meaningless)
1049 :     \dQuote{head of neighbourhood} is no longer \code{show}n.
1050 :    
1051 :     \item The \code{"sts"} class now has a \code{dimnames}-method
1052 :     instead of a \code{colnames}-method. Furthermore, the redundant
1053 :     \code{nrow} and \code{ncol} methods have been removed
1054 :     (the \code{dim}-method is sufficient).
1055 :    
1056 :     \item If a \code{map} is provided when \code{initialize()}ing an
1057 :     \code{"sts"} object, it is now verified that all \code{observed}
1058 :     regions are part of the \code{map} (matched by \code{row.names}).
1059 :    
1060 :     \item In \code{stsplot_space()}, extra (unobserved) regions of the
1061 :     \code{map} are no longer dropped but shown with a dashed border by
1062 :     default.
1063 :     }
1064 :     }
1065 :    
1066 :     }
1067 :    
1068 :    
1069 :    
1070 :     \section{Changes in surveillance version 1.8-1 (2014-10-29)}{
1071 :    
1072 :     \subsection{NEW FEATURES}{
1073 :     \itemize{
1074 :     \item The \code{R0}-method for \code{"twinstim"} gained an argument
1075 :     \code{newcoef} to simplify computation of reproduction numbers with
1076 :     a different parameter vector (also used for Monte Carlo CI's).
1077 :    
1078 :     \item New plot \code{type="neweights"} for \code{"hhh4"} fits.
1079 :    
1080 :     \item The \code{scores()} function allows the selection of multiple
1081 :     \code{units} (by index or name) for which to compute (averaged) proper
1082 :     scores. Furthermore, one can now select \code{which} scores to compute.
1083 :    
1084 :     \item Added a \code{formula}-method for \code{"hhh4"} fits to
1085 :     extract the \code{f} specifications of the three components from the
1086 :     control list.
1087 :    
1088 :     \item The \code{update()}-method for fitted \code{"hhh4"} models
1089 :     gained an argument \code{S} for convenient modification of component
1090 :     seasonality using \code{addSeason2formula()}.
1091 :    
1092 :     \item The new auxiliary function \code{layout.labels()} generates an
1093 :     \code{sp.layout} item for \code{spplot()} in order to draw labels.
1094 :    
1095 :     \item When generating the \code{pit()} histogram with a single
1096 :     predictive CDF \code{pdistr}, the \code{\dots} arguments can now be
1097 :     \code{x}-specific and are recycled if necessary using \code{mapply()}.
1098 :     If \code{pdistr} is a list of CDFs, \code{pit()} no longer requires
1099 :     the functions to be vectorized.
1100 :    
1101 :     \item New method \code{as.epidata.data.frame()}, which constructs the
1102 :     start/stop SIR event history format from a simple individual-based
1103 :     data frame (e.g., \code{hagelloch.df}).
1104 :    
1105 :     \item New argument \code{w} in \code{as.epidata.default()} to
1106 :     generate covariate-based weights for the force of infection in
1107 :     \code{twinSIR}. The \code{f} argument is for distance-based weights.
1108 :    
1109 :     \item The result of \code{profile.twinSIR()} gained a class and an
1110 :     associated \code{plot}-method.
1111 :     }
1112 :     }
1113 :    
1114 :     \subsection{MAJOR CHANGES}{
1115 :     \itemize{
1116 :     \item For multivariate \code{oneStepAhead()} predictions,
1117 :     \code{scores(..., individual=TRUE)} now returns a 3d array instead
1118 :     of a collapsed matrix. Furthermore, the scores computed by default
1119 :     are \code{c("logs","rps","dss","ses")}, excluding the normalized
1120 :     squared error score \code{"nses"} which is improper.
1121 :    
1122 :     \item The plot-\code{type="season"} for \code{"hhh4"} fits now by
1123 :     default plots the multiplicative effect of seasonality on the
1124 :     respective component (new argument \code{intercept=FALSE}).
1125 :     The default set of components to plot has also changed.
1126 :    
1127 :     \item When \code{as.epidata()} and \code{simEpidata()} calculate
1128 :     distance-based epidemic weights from the \code{f} functions, they no
1129 :     longer set the distance of an infectious individual to itself
1130 :     artificially to \code{Inf}.
1131 :     This changes the corresponding columns in the \code{"epidata"} in
1132 :     rows of currently infectious individuals, but the \code{twinSIR}
1133 :     model itself is invariant, since only rows with \code{atRiskY=1}
1134 :     contribute to the likelihood.
1135 :    
1136 :     \item Several modifications and corrections in \code{data("hagelloch")}.
1137 :     }
1138 :     }
1139 :    
1140 :     \subsection{MINOR CHANGES}{
1141 :     \itemize{
1142 :     \item Better plotting of \code{stsNC} objects by writing an own plot
1143 :     method for them. Prediction intervals are now shown jointly with the
1144 :     point estimate.
1145 :    
1146 :     \item Reduced package size by applying \code{tools::resaveRdaFiles}
1147 :     to some large datasets and by building the package with
1148 :     \code{--compact-vignettes=both}, i.e., using additional GhostScript
1149 :     compression with ebook quality, see \code{?tools::compactPDF}.
1150 :    
1151 :     \item Added \code{units} argument to \code{stsplot_time} to select
1152 :     only a subset of the multivariate time series for plotting.
1153 :    
1154 :     \item The \code{untie}-method for class \code{"epidataCS"} gained an
1155 :     argument \code{verbose} which is now \code{FALSE} by default.
1156 :    
1157 :     \item \code{"epidataCS"} objects store the \code{clipper} used
1158 :     during generation as attribute of \code{$events$.influenceRegion}.
1159 :    
1160 :     \item In \code{plotHHH4_fitted()}, the argument \code{legend.observed}
1161 :     now defaults to \code{FALSE}.
1162 :    
1163 :     \item The default weights for the spatio-temporal component in
1164 :     \code{hhh4} models now are \code{neighbourhood(stsObj) == 1}.
1165 :     The previous default \code{neighbourhood(stsObj)} does not make
1166 :     sense for the newly supported \code{nbOrder} neighbourhood matrices
1167 :     (shortest-path distances). The new default makes no difference for
1168 :     (old) models with binary adjacency matrices in the neighbourhood
1169 :     slot of the \code{stsObj}.
1170 :    
1171 :     \item The default for nonparametric weights \code{W_np()} in
1172 :     \code{hhh4()} is now to assume zero weight for neighbourhood orders
1173 :     above \code{maxlag}, i.e., \code{W_np()}'s argument \code{to0} now
1174 :     defaults to \code{TRUE}.
1175 :    
1176 :     \item Added a \code{verbose} argument to \code{permutationTest()},
1177 :     which defaults to \code{FALSE}. The previous behaviour corresponds
1178 :     to \code{verbose=TRUE}.
1179 :    
1180 :     \item \code{simulate.twinstim()} now by default uses the original
1181 :     \code{data$W} as observation region.
1182 :    
1183 :     \item The \code{data("measlesWeserEms")} contain two additional
1184 :     variables in the \code{@map@data} slot: \code{"vaccdoc.2004"} and
1185 :     \code{"vacc1.2004"}.
1186 :    
1187 :     \item The plot-method for \code{"epidata"} objects now uses colored
1188 :     lines by default.
1189 :    
1190 :     \item The \pkg{surveillance} package now depends on \R >= 3.0.2,
1191 :     which, effectively, is the minimum version required since
1192 :     \pkg{surveillance} 1.7-0 (see the corresponding NEWS below).
1193 :    
1194 :     \item The two diagnostic plots of \code{checkResidualProcess()} are
1195 :     now by default plotted side by side (\code{mfrow=c(1,2)}) instead of
1196 :     one below the other.
1197 :     }
1198 :     }
1199 :    
1200 :     \subsection{BUG FIXES}{
1201 :     \itemize{
1202 :     \item In \code{farringtonFlexible} alarms are now for
1203 :     \code{observed>upperbound} and not for \code{observed>=upperbound}
1204 :     which was not correct.
1205 :    
1206 :     \item Fixed duplicate \code{"functions"} element resulting from
1207 :     \code{update.twinstim(*,model=TRUE)} and ensured that
1208 :     \code{"twinstim"} objects always have the same components (some may
1209 :     be \code{NULL}).
1210 :    
1211 :     \item \code{animate.epidata} works again with the
1212 :     \CRANpkg{animation} package (\code{ani.options("outdir")} was
1213 :     removed in version 2.3)
1214 :    
1215 :     \item For \code{hhh4} models with random effects, \code{confint()}
1216 :     only worked if argument \code{parm} was specified.
1217 :    
1218 :     \item Computing one-sided AIC weights by simulation for \code{twinSIR}
1219 :     models with more than 2 epidemic covariates now is more robust (by
1220 :     rescaling the objective function for the quadratic programming
1221 :     solver) and twice as fast (due to code optimization).
1222 :    
1223 :     \item \code{simulate.twinstim(..., rmarks=NULL)} can now handle the
1224 :     case where \code{data} has no events within the simulation period
1225 :     (by sampling marks from all of \code{data$events}).
1226 :    
1227 :     \item The \code{lambda.h} values of simulated events in
1228 :     \code{"simEpidataCS"} objects were wrong if the model contained an
1229 :     endemic intercept (which is usually the case).
1230 :    
1231 :     \item Automatic choice of color breaks in the \code{animate}-method
1232 :     for class \code{"sts"} now also works for incidence maps (i.e., with
1233 :     a \code{population} argument).
1234 :    
1235 :     \item \code{hhh4()} did not allow the use of nonparametric
1236 :     neighbourhood weights \code{W_np()} with \code{maxlag=2}.
1237 :    
1238 :     \item \code{scores()} did not work for multivariate \code{oneStepAhead()}
1239 :     predictions if both \code{individual=TRUE} and \code{sign=TRUE}, and
1240 :     it could not handle a \code{oneStepAhead()} prediction of only one
1241 :     time point. Furthermore, the \code{"sign"} column of
1242 :     \code{scores(..., sign=TRUE)} was wrong (reversed).
1243 :    
1244 :     \item For \code{"epidataCS"} with only one event,
1245 :     \code{epidataCSplot_space()} did not draw the point.
1246 :    
1247 :     \item The trivial (identity) call
1248 :     \code{aggregate(stsObj, nfreq=stsObj@freq)} did not work.
1249 :     }
1250 :     }
1251 :    
1252 :     }
1253 :    
1254 :    
1255 :    
1256 :     \section{Changes in surveillance version 1.8-0 (2014-06-16)}{
1257 :    
1258 :     \subsection{PACKAGE INFRASTRUCTURE}{
1259 :     \itemize{
1260 :     \item Package \pkg{surveillance} now depends on newer versions of
1261 :     packages \CRANpkg{sp} (>= 1.0-15), \CRANpkg{polyCub} (>= 0.4-2),
1262 :     and \CRANpkg{spatstat} (>= 1.36-0).
1263 :     The \R packages \pkg{INLA} and \CRANpkg{runjags} are now suggested
1264 :     to support a new outbreak detection algorithm (\code{boda()}) and
1265 :     the new \code{nowcast()}ing procedure, respectively.
1266 :     The \R packages for \CRANpkg{lattice}, \CRANpkg{grid},
1267 :     \CRANpkg{gridExtra}, and \CRANpkg{scales} are suggested for
1268 :     added visualization facilities.
1269 :    
1270 :     \item More tests have been implemented to ensure package integrity.
1271 :     We now use \CRANpkg{testthat} instead of the outdated package
1272 :     \CRANpkg{RUnit}.
1273 :    
1274 :     \item \code{hhh4()} fits now have class \code{"hhh4"} instead of
1275 :     \code{"ah4"}, for consistency with \code{twinstim()},
1276 :     \code{twinSIR()}, and to follow the common convention (cp. \code{lm()}).
1277 :     Standard S3-methods for the old \code{"ah4"} name are still
1278 :     available for backwards compatibility but may be removed in the
1279 :     future.
1280 :    
1281 :     \item Plot variants for \code{"sts"} objects have been cleaned up:
1282 :     The functions implementing the various plot types
1283 :     (\code{stsplot_*}, previously named \code{plot.sts.*})
1284 :     are now exported and documented separately.
1285 :     }
1286 :     }
1287 :    
1288 :     \subsection{NEW FEATURES}{
1289 :     \itemize{
1290 :     \item The \code{nowcast} procedure has been completely re-written to
1291 :     handle the inherit right-truncation of reporting data (best
1292 :     visualized as a reporting triangle). The new code implements the
1293 :     generalized-Dirichlet and the hierarchical Bayesian approach described in
1294 :     \enc{Höhle}{Hoehle} and an der Heiden (2014). No backwards
1295 :     compatibility to the old nowcasting procedure is given.
1296 :    
1297 :     \item The package contains a new monitoring function
1298 :     \code{boda}. This is a first experimental surveillance
1299 :     implementation of the Bayesian Outbreak Detection Algorithm (BODA)
1300 :     proposed in Manitz and \enc{Höhle}{Hoehle} (2012). The function
1301 :     relies on the non-CRAN package \pkg{INLA}, which has to be installed first
1302 :     in order to use this function. Expect initial problems.
1303 :    
1304 :    
1305 :     \item New \code{toLatex}-method for \code{"sts"} objects.
1306 :    
1307 :     \item The new function \code{stsplot_space()} provides an improved
1308 :     map plot of disease incidence for \code{"sts"} objects aggregated
1309 :     over time. It corresponds to the new \code{type = observed ~ unit}
1310 :     of the \code{stsplot}-method, and supersedes
1311 :     \code{type = observed ~ 1|unit} (except for alarm shading).
1312 :    
1313 :     \item An \code{animate()}-method for the \code{"sts"} class provides
1314 :     a new implementation for animated maps (superseding the \code{plot}
1315 :     \code{type=observed ~ 1 | unit * time}) with an optional evolving
1316 :     time series plot below the map.
1317 :    
1318 :     \item The \code{plot()} method for \code{"sts"} objects with epochs as
1319 :     dates is now made more flexible by introducing the arguments
1320 :     \code{xaxis.tickFreq}, \code{xaxis.labelFreq} and
1321 :     \code{xaxis.labelFormat}. These allow the specification of
1322 :     tick-marks and labelling based on \code{strftime} compatible
1323 :     conversion codes -- independently if data are daily, weekly, monthly,
1324 :     etc. As a consequence, the old argument \code{xaxis.years} is
1325 :     removed. See \code{stsplot_time()} for more information.
1326 :    
1327 :    
1328 :     \item Inference for neighbourhood weights in \code{hhh4()} models:
1329 :     \code{W_powerlaw()} and \code{W_np()} both implement weights
1330 :     depending on the order of neighbourhood between regions, a power-law
1331 :     decay and nonparametric weights, i.e., unconstrained estimation of
1332 :     individual weights for each neighbourhood order.
1333 :    
1334 :     \item \code{hhh4()} now allows the inclusion of multiplicative
1335 :     offsets also in the epidemic components \code{"ar"} and \code{"ne"}.
1336 :    
1337 :     \item \code{hhh4()} now has support for \code{lag != 1} in the
1338 :     autoregressive and neighbor-driven components. The applied lags are
1339 :     stored as component \code{"lags"} of the return value (previously
1340 :     there was an unused component \code{"lag"} which was always 1 and
1341 :     has been removed now).
1342 :    
1343 :     \item \code{oneStepAhead()}:
1344 :     \itemize{
1345 :     \item Added support for parallel computation of predictions using
1346 :     \code{mclapply()} from package \pkg{parallel}.
1347 :     \item New argument \code{type} with a new \code{type}
1348 :     \code{"first"} to base all subsequent one-step-ahead predictions
1349 :     on a single initial fit.
1350 :     \item Nicer interpretation of \code{verbose} levels, and
1351 :     \code{txtProgressBar()}.
1352 :     }
1353 :    
1354 :     \item The \code{plot()}-method for fitted \code{hhh4()} objects now
1355 :     offers three additional types of plots: component seasonality,
1356 :     seasonal or time course of the dominant eigenvalue, and maps
1357 :     of estimated random intercepts. It is documented and more customizable.
1358 :     Note that argument order and some names have changed:
1359 :     \code{i} -> \code{units}, \code{title} -> \code{names}.
1360 :    
1361 :     \item (Deviance) \code{residuals()}-method for fitted \code{hhh4()}
1362 :     models.
1363 :    
1364 :     \item Added methods of \code{vcov()} and \code{nobs()}
1365 :     for the \code{"hhh4"} class. For \code{AIC()} and \code{BIC()}, the
1366 :     default methods work smoothly now (due to changes to
1367 :     \code{logLik.hhh4()} documented below).
1368 :    
1369 :    
1370 :     \item New predefined interaction functions for \code{twinstim()}:
1371 :     \code{siaf.student()} implements a \eqn{t}-kernel for the distance
1372 :     decay, and \code{siaf.step()} and \code{tiaf.step()} provide step
1373 :     function kernels (which may also be invoked by specifying the
1374 :     vector of knots as the \code{siaf} or \code{tiaf} argument in
1375 :     \code{twinstim}).
1376 :    
1377 :     \item Numerical integration over polygonal domains in the \code{F}
1378 :     and \code{Deriv} components of \code{siaf.powerlaw()} and
1379 :     \code{siaf.powerlawL()} is much faster and more accurate now since
1380 :     we use the new \code{polyCub.iso()} instead of \code{polyCub.SV()}
1381 :     from package \CRANpkg{polyCub}.
1382 :    
1383 :     \item New \code{as.stepfun()}-method for \code{"epidataCS"} objects.
1384 :    
1385 :     \item \code{plot.epidataCS()}:
1386 :     \itemize{
1387 :     \item The spatial plot has new arguments to automatically add
1388 :     legends to the plot: \code{legend.types} and \code{legend.counts}.
1389 :     It also gained an \code{add} argument.
1390 :     \item The temporal plot now supports type-specific sub-histograms,
1391 :     additional lines for the cumulative number of events, and an
1392 :     automatic legend.
1393 :     }
1394 :    
1395 :     \item The new function \code{glm_epidataCS()} can be used to fit
1396 :     an endemic-only \code{twinstim()} via \code{glm()}.
1397 :     This is mainly provided for testing purposes since wrapping into
1398 :     \code{glm} usually takes longer.
1399 :     }
1400 :     }
1401 :    
1402 :     \subsection{MAJOR CHANGES}{
1403 :     \itemize{
1404 :     \item Fitted \code{hhh4()} objects no longer contain the associated
1405 :     \code{"sts"} data twice: it is now only stored as \code{$stsObj}
1406 :     component, the hidden duplicate in \code{$control$data$.sts} was
1407 :     dropped, which makes fitted objects substantially smaller.
1408 :    
1409 :     \item \code{logLik.hhh4()} always returns an object of class
1410 :     \code{"logLik"} now; for random effects models, its \code{"df"}
1411 :     attribute is \code{NA_real_}. Furthermore, for non-convergent fits,
1412 :     \code{logLik.hhh4()} gives a warning and returns \code{NA_real_};
1413 :     previously, an error was thrown in this case.
1414 :    
1415 :     \item \code{oneStepAhead()}:
1416 :     \itemize{
1417 :     \item Default of \code{tp[2]} is now the penultimate time point of
1418 :     the fitted subset (not of the whole \code{stsObj}).
1419 :     \item \code{+1} on rownames of \code{$pred} (now the same as for
1420 :     \code{$observed}).
1421 :     }
1422 :    
1423 :     \item The optional \code{"twinstim"} result components
1424 :     \code{fisherinfo}, \code{tau}, and \code{functions} are always
1425 :     included. They are set to \code{NULL} if they are not applicable
1426 :     instead of missing completely (as before), such that all
1427 :     \code{"twinstim"} objects have the same list structure.
1428 :    
1429 :     \item \code{iafplot()} ...
1430 :     \itemize{
1431 :     \item invisibly returns a matrix containing the plotted
1432 :     values of the (scaled) interaction function (and the confidence
1433 :     interval as an attribute).
1434 :     Previously, nothing (\code{NULL}) was returned.
1435 :     \item detects a type-specific interaction function and by default
1436 :     uses \code{types=1} if it is not type-specific.
1437 :     \item has better default axis ranges.
1438 :     \item adapts to the new step function kernels (with new arguments
1439 :     \code{verticals} and \code{do.points}).
1440 :     \item supports logarithmic axes (via new \code{log} argument
1441 :     passed on to \code{plot.default}).
1442 :     \item optionally respects \code{eps.s} and \code{eps.t},
1443 :     respectively (by the new argument \code{truncated}).
1444 :     \item now uses \code{scaled=TRUE} by default.
1445 :     }
1446 :    
1447 :     \item The argument \code{colTypes} of
1448 :     \code{plot.epidataCS(,aggregate="space")} is deprecated (use
1449 :     \code{points.args$col} instead).
1450 :    
1451 :     \item The events in an \code{"epidataCS"} object no longer have
1452 :     a reserved \code{"ID"} column.
1453 :     }
1454 :     }
1455 :    
1456 :     \subsection{MINOR CHANGES}{
1457 :     \itemize{
1458 :     \item \code{hhh4()} now stores the runtime just like \code{twinstim()}.
1459 :     \item Take \code{verbose=FALSE} in \code{hhh4()} more seriously.
1460 :     \item \code{hhh4()} issues a \code{warning()} if non-convergent.
1461 :     \item The following components of a \code{hhh4()} fit now have names:
1462 :     \code{"se"}, \code{"cov"}, \code{"Sigma"}.
1463 :     \item The new default for \code{pit()} is to produce the plot.
1464 :    
1465 :     \item The \code{twinstim()} argument \code{cumCIF} now defaults to
1466 :     \code{FALSE}.
1467 :     \item \code{update.twinstim()} no longer uses recursive
1468 :     \code{modifyList()} for the \code{control.siaf} argument. Instead,
1469 :     the supplied new list elements (\code{"F"}, \code{"Deriv"})
1470 :     completely replace the respective elements from the original
1471 :     \code{control.siaf} specification.
1472 :     \item \code{siaf.lomax()} is now defunct (it has been deprecated
1473 :     since version 1.5-2); use \code{siaf.powerlaw()} instead.
1474 :     \item Allow the default \code{adapt}ive bandwidth to be specified via the
1475 :     \code{F.adaptive} argument in \code{siaf.gaussian()}.
1476 :     \item Unsupported options (\code{logpars=FALSE},
1477 :     \code{effRangeProb}) have been dropped from \code{siaf.powerlaw()}
1478 :     and \code{siaf.powerlawL()}.
1479 :     \item More rigorous checking of \code{tiles} in
1480 :     \code{simulate.twinstim()} and \code{intensityplot.twinstim}.
1481 :    
1482 :     \item \code{as.epidataCS()} gained a \code{verbose} argument.
1483 :     \item \code{animate.epidataCS()} now by default does not draw
1484 :     influence regions (\code{col.influence=NULL}), is \code{verbose} if
1485 :     \code{interactive()}, and ignores \code{sleep} on non-interactive
1486 :     devices.
1487 :    
1488 :     \item The \code{multiplicity}-generic and its default method have
1489 :     been integrated into \CRANpkg{spatstat} and are imported from there.
1490 :     }
1491 :     }
1492 :    
1493 :     \subsection{DATA}{
1494 :     \itemize{
1495 :     \item The polygon representation of Germany's districts (
1496 :     \code{system.file("shapes", "districtsD.RData", package="surveillance")}
1497 :     ) has been simplified further. The union of \code{districtsD} is
1498 :     used as observation window \code{W} in \code{data("imdepi")}. The
1499 :     exemplary \code{twinstim()} fit \code{data("imdepifit")} has been
1500 :     updated as well. Furthermore, \code{row.names(imdepi$events)} have
1501 :     been reset (chronological index), and numerical differences
1502 :     in \code{imdepi$events$.influenceRegion} are due to changes in
1503 :     \CRANpkg{polyclip} 1.3-0.
1504 :    
1505 :     \item The Campylobacteriosis data set \code{campyDE}, where absolute
1506 :     humidity is used as concurrent covariate to adjust the outbreak
1507 :     detection is added to the package to exemplify \code{boda()}.
1508 :    
1509 :     \item New \code{data("measlesWeserEms")} (of class \code{"sts"}),
1510 :     a corrected version of \code{data("measles.weser")} (of the old
1511 :     \code{"disProg"} class).
1512 :     }
1513 :     }
1514 :    
1515 :     \subsection{BUG FIXES}{
1516 :     \itemize{
1517 :     \item Fixed a bug in \code{LRCUSUM.runlength} where computations
1518 :     were erroneously always done under the in-control parameter
1519 :     \code{mu0} instead of \code{mu}.
1520 :    
1521 :     \item Fixed a bug during alarm plots (\code{stsplot_alarm()}),
1522 :     where the use of \code{alarm.symbol} was ignored.
1523 :    
1524 :     \item Fixed a bug in \code{algo.glrnb} where the overdispersion
1525 :     parameter \code{alpha} from the automatically fitted \code{glm.nb}
1526 :     model (fitted by \code{estimateGLRNbHook}) was incorrectly taken as
1527 :     \code{mod[[1]]$theta} instead of \code{1/mod[[1]]$theta}. The error is
1528 :     due to a different parametrization of the negative binomial
1529 :     distribution compared to the parametrization in \enc{Höhle}{Hoehle}
1530 :     and Paul (2008).
1531 :    
1532 :     \item The score function of \code{hhh4()} was wrong when fitting
1533 :     endemic-only models to a \code{subset} including the first time
1534 :     point. This led to \dQuote{false convergence}.
1535 :    
1536 :     \item \code{twinstim()} did not work without an endemic offset if
1537 :     \code{is.null(optim.args$par)}.
1538 :     }
1539 :     }
1540 :    
1541 :     }
1542 :    
1543 :    
1544 :    
1545 :     \section{Changes in surveillance version 1.7-0 (2013-11-19)}{
1546 :    
1547 :     \subsection{SYNOPSIS}{
1548 :     \itemize{
1549 :     \item Package \CRANpkg{gpclib} is no longer necessary for the
1550 :     construction of \code{"epidataCS"}-objects. Instead, we make use of
1551 :     the new dedicated package \CRANpkg{polyclip} (licensed under the
1552 :     BSL) for polygon clipping operations (via
1553 :     \code{spatstat::intersect.owin()}). This results in a slightly
1554 :     different \code{$events$.influenceRegion} component of
1555 :     \code{"epidataCS"} objects, one reason being that
1556 :     \pkg{polyclip} uses integer arithmetic.
1557 :     Change of \code{twinstim()} estimates for a newly created
1558 :     \code{"epidataCS"} compared with the same data prepared in earlier
1559 :     versions should be very small (e.g., for \code{data("imdepifit")}
1560 :     the mean relative difference of coefficients is 3.7e-08, while the
1561 :     \code{logLik()} is \code{all.equal()}).
1562 :     As an alternative, \pkg{rgeos} can still be chosen to do the polygon
1563 :     operations.
1564 :    
1565 :     \item The \pkg{surveillance}-internal code now depends on
1566 :     \R >= 2.15.2 (for \code{nlminb()} \code{NA} fix of PR#15052,
1567 :     consistent \code{rownames(model.matrix)} of PR#14992,
1568 :     \code{paste0()}, \code{parallel::mcmapply()}). However, the
1569 :     required recent version of \pkg{spatstat} (1.34-0, for
1570 :     \pkg{polyclip}) actually needs \R >= 3.0.2, which therefore also
1571 :     applies to \pkg{surveillance}.
1572 :    
1573 :     \item Some minor new features and changes are documented below.
1574 :     }
1575 :     }
1576 :    
1577 :     \subsection{NEW FEATURES}{
1578 :     \itemize{
1579 :     \item Functions \code{unionSpatialPolygons()} and
1580 :     \code{intersectPolyCircle()} are now exported. Both are wrappers
1581 :     around functionality from different packages supporting polygon
1582 :     operations: for determining the union of all subpolygons of a
1583 :     \code{"SpatialPolygons"} object, and the intersection of a polygonal
1584 :     and a circular domain, respectively.
1585 :     \item \code{discpoly()} moved back from \CRANpkg{polyCub}
1586 :     to \pkg{surveillance}.
1587 :     }
1588 :     }
1589 :    
1590 :     \subsection{MINOR CHANGES}{
1591 :     \itemize{
1592 :     \item \pkg{surveillance} now Depends on \CRANpkg{polyCub} (>= 0.4-0)
1593 :     and not only Imports it (which avoids \code{::}-references in
1594 :     .GlobalEnv-made functions).
1595 :     \item Nicer default axis labels for \code{iafplot()}.
1596 :     \item For \code{twinstim()}, the default is now to \code{trace}
1597 :     every iteration instead of every fifth only.
1598 :     \item Slightly changed default arguments for \code{plot.epidata()}:
1599 :     \code{lwd} (1->2), \code{rug.opts} (\code{col} is set according to
1600 :     \code{which.rug})
1601 :     \item \code{twinstim()} saves the vector of \code{fixed}
1602 :     coefficients as part of the returned \code{optim.args} component,
1603 :     such that these will again be held fixed upon \code{update()}.
1604 :     \item The \code{plot}-method for \code{hhh4()}-fits allows for
1605 :     region selection by name.
1606 :     }
1607 :     }
1608 :    
1609 :     }
1610 :    
1611 :    
1612 :    
1613 :     \section{Changes in surveillance version 1.6-0 (2013-09-03)}{
1614 :    
1615 :     \subsection{SYNOPSIS}{
1616 :     \itemize{
1617 :     \item The \code{polyCub}-methods for cubature over polygonal domains
1618 :     have been moved to the new dedicated package \CRANpkg{polyCub},
1619 :     since they are of a rather general use. The \code{discpoly()} function
1620 :     has also been moved to that package.
1621 :    
1622 :     \item As a replacement for the license-restricted \pkg{gpclib} package,
1623 :     the \pkg{rgeos} package is now used by default
1624 :     (\code{surveillance.options(gpclib=FALSE)}) in generating
1625 :     \code{"epidataCS"} (polygon intersections, slightly slower).
1626 :     Therefore, when installing \pkg{surveillance} version 1.6-0, the
1627 :     system requirements for \CRANpkg{rgeos} have to be met, i.e., GEOS
1628 :     must be available on the system. On Linux variants this means
1629 :     installing \file{libgeos} (\file{libgeos-dev}).
1630 :    
1631 :     \item The improved Farrington method described in Noufaily et al.
1632 :     (2012) is now available as function \code{farringtonFlexible()}.
1633 :    
1634 :     \item New handling of reference dates in \code{algo.farrington()} for
1635 :     \code{"sts"} objects with \code{epochAsDate=TRUE}. Instead of always
1636 :     going back in time to the next Date in the \code{"epoch"} slot, the
1637 :     function now determines the \emph{closest} Date. Note that this
1638 :     might lead to slightly different results for the upperbound
1639 :     compared to previously. Furthermore, the functionality is only
1640 :     tested for weekly data (monthly data are experimental). The same
1641 :     functionality applies to \code{farringtonFlexible()}.
1642 :    
1643 :     \item To make the different retrospective modelling frameworks of
1644 :     the \pkg{surveillance} package jointly applicable, it is now possible
1645 :     to convert (aggregate) \code{"epidataCS"}
1646 :     (continuous-time continuous-space data) into an \code{"sts"} object
1647 :     (multivariate time series of counts) by the new function
1648 :     \code{epidataCS2sts}.
1649 :    
1650 :     \item Simulation from \code{hhh4} models has
1651 :     been re-implemented, which fixes a bug and makes it more flexible
1652 :     and compatible with a wider class of models.
1653 :    
1654 :     \item The \code{map}-slot of the \code{"sts"} class now requires
1655 :     \code{"SpatialPolygons"} (only) instead of
1656 :     \code{"SpatialPolygonsDataFrame"}.
1657 :    
1658 :     \item Re-implementation of \code{oneStepAhead()} for
1659 :     \code{hhh4}-models with a bug fix, some speed-up and more options.
1660 :    
1661 :     \item Slight speed-up for \code{hhh4()} fits,
1662 :     e.g., by more use of \code{.rowSums()} and \code{.colSums()}.
1663 :    
1664 :     \item Crucial speed-up for \code{twinstim()} fits by more efficient
1665 :     code: \code{mapply}, dropped clumsy \code{for}-loop in
1666 :     \code{fisherinfo}, new argument \code{cores} for parallel
1667 :     computing via forking (not available on Windows).
1668 :    
1669 :     \item Some further new features, minor changes, and bug fixes are
1670 :     described in the following subsections.
1671 :     }
1672 :     }
1673 :    
1674 :     \subsection{NEW FEATURES}{
1675 :     \itemize{
1676 :     \item Using \code{tiaf.exponential()} in a \code{twinstim()} now works
1677 :     with \code{nTypes=1} for multi-type data.
1678 :    
1679 :     \item A legend can be added automatically in \code{iafplot()}.
1680 :    
1681 :     \item The \code{untie} methods are now able to produce jittered points
1682 :     with a required minimum separation (\code{minsep}).
1683 :    
1684 :     \item \code{simulate.ah4} gained a \code{simplify} argument.
1685 :    
1686 :     \item New \code{update}-method for fitted \code{hhh4}-models
1687 :     (class \code{"ah4"}).
1688 :    
1689 :     \item \code{oneStepAhead()} has more options: specify time range
1690 :     (not only start), choose type of start values, \code{verbose}
1691 :     argument.
1692 :    
1693 :     \item \code{pit()} allows for a list of predictive distributions
1694 :     (\code{pdistr}), one for each observation \code{x}.
1695 :    
1696 :     \item New spatial auxiliary function \code{polyAtBorder()}
1697 :     indicating polygons at the border (for a \code{"SpatialPolygons"}
1698 :     object).
1699 :    
1700 :     \item \code{animate.epidataCS()} allows for a \code{main} title and
1701 :     can show a progress bar.
1702 :     }
1703 :     }
1704 :    
1705 :     \subsection{MINOR CHANGES}{
1706 :     \itemize{
1707 :     \item Changed parametrization of \code{zetaweights()} and completed
1708 :     its documentation (now no longer marked as experimental).
1709 :    
1710 :     \item \code{twinstim(...)$converged} is \code{TRUE} if
1711 :     the optimization routine converged (as before) but contains
1712 :     the failure message otherwise.
1713 :    
1714 :     \item Increased default \code{maxit} for the Nelder-Mead optimizer
1715 :     in \code{hhh4} from 50 to 300, and removed default artificial lower
1716 :     bound (-20) on intercepts of epidemic components.
1717 :    
1718 :     \item Renamed returned list from \code{oneStepAhead}
1719 :     (mean->pred, x->observed, params->coefficients,
1720 :     variances->Sigma.orig) for consistency, and
1721 :     \code{oneStepAhead()$psi} is only non-\code{NULL} if we have a
1722 :     NegBin model.
1723 :    
1724 :     \item Argument order of \code{pit()} has changed, which is also
1725 :     faster now and got additional arguments \code{relative} and
1726 :     \code{plot}.
1727 :    
1728 :     \item \code{twinstim(...)$runtime} now contains the complete
1729 :     information from \code{proc.time()}.
1730 :     }
1731 :     }
1732 :    
1733 :     \subsection{BUG FIXES}{
1734 :     \itemize{
1735 :     \item Fixed a bug in function
1736 :     \code{refvalIdxByDate()} which produced empty reference values
1737 :     (i.e. \code{NA}s) in case the Date entries of \code{epoch} were not
1738 :     mondays. Note: The function works by subtracting \code{1:b} years from the
1739 :     date of the range value and then takes the span \code{-w:w} around this
1740 :     value. For each value in this set it is determined whether the
1741 :     closest date in the epoch slot is obtained by going forward or
1742 :     backward. Note that this behaviour is now slightly changed compared
1743 :     to previously, where we \emph{always} went back in time.
1744 :    
1745 :     \item \code{algo.farrington()}: Reference values too far back in time
1746 :     and hence not being in the \code{"epoch"} slot of the \code{"sts"}
1747 :     object are now ignored (previously the resulting \code{NA}s caused the
1748 :     function to halt). A warning is displayed in this case.
1749 :    
1750 :     \item \code{hhh4}: The entry \eqn{(5,6)} of the marginal Fisher
1751 :     information matrix in models with random intercepts in all three
1752 :     components was incorrect.
1753 :     If \code{nlminb} was used as optimizer for the variance parameters
1754 :     (using the negative marginal Fisher information as Hessian), this
1755 :     could have caused false convergence (with warning) or minimally
1756 :     biased convergence (without warning).
1757 :     As a consequence, the \code{"Sigma.cov"} component of the
1758 :     \code{hhh4()} result, which is the inverse of the marginal Fisher
1759 :     information matrix at the MLE, was also wrong.
1760 :    
1761 :     \item \code{untie.matrix()} could have produced jittering greater than
1762 :     the specified \code{amount}.
1763 :    
1764 :     \item \code{hhh4}: if there are no random intercepts, the
1765 :     redundant \code{updateVariance} steps are no longer evaluated.
1766 :    
1767 :     \item \code{update.twinstim()} did not work with
1768 :     \code{optim.args=..1} (e.g., if updating a list of models with lapply).
1769 :     Furthermore, if adding the \code{model} component only, the
1770 :     \code{control.siaf} and \code{optim.args} components were lost.
1771 :    
1772 :     \item \code{earsC} should now also work with multivariate
1773 :     \code{sts} time-series objects.
1774 :    
1775 :     \item The last week in \code{data(fluBYBW)} (row index 417) has been
1776 :     removed. It corresponded to week 1 in year 2009 and was wrong
1777 :     (an artifact, filled with zero counts only).
1778 :     Furthermore, the regions in \code{@map} are now ordered the same as
1779 :     in \code{@observed}.
1780 :    
1781 :     \item Fixed start value of the overdispersion parameter in
1782 :     \code{oneStepAhead} (must be on internal log-scale, not
1783 :     reparametrized as returned by \code{coef()} by default).
1784 :    
1785 :     \item When subsetting \code{"sts"} objects in time, \code{@start} was
1786 :     updated but not \code{@epoch}.
1787 :    
1788 :     \item \code{pit} gave \code{NA} results if any \code{x[-1]==0}.
1789 :    
1790 :     \item The returned \code{optim.args$par} vector in \code{twinstim()}
1791 :     was missing any fixed parameters.
1792 :    
1793 :     \item \code{hhh4()} did not work with time-varying neighbourhood
1794 :     weights due to an error in the internal \code{checkWeightsArray()}
1795 :     function.
1796 :     }
1797 :     }
1798 :    
1799 :     }
1800 :    
1801 :    
1802 :    
1803 :     \section{Changes in surveillance version 1.5-4 (2013-04-21)}{
1804 :    
1805 :     \subsection{SYNOPSIS}{
1806 :     \itemize{
1807 :     \item Fixed obsolete \code{.path.package()} calls.
1808 :     \item Small corrections in the documentation.
1809 :     \item \code{update.twinstim()} performs better in preserving
1810 :     the original initial values of the parameters.
1811 :     \item New pre-defined spatial interaction function
1812 :     \code{siaf.powerlawL()}, which implements a _L_agged power-law
1813 :     kernel, i.e. accounts for uniform short-range dispersal.
1814 :     }
1815 :     }
1816 :    
1817 :     }
1818 :    
1819 :    
1820 :    
1821 :     \section{Changes in surveillance version 1.5-2 (2013-03-15)}{
1822 :    
1823 :     \subsection{SYNOPSIS}{
1824 :     \itemize{
1825 :     \item New method for outbreak detection: \code{earsC}
1826 :     (CUSUM-method described in the CDC Early Aberration Reporting
1827 :     System, see Hutwagner et al, 2003).
1828 :     \item New features and minor bug fixes for the "\code{twinstim}"
1829 :     part of the package (see below).
1830 :     \item Yet another p-value formatting function \code{formatPval()}
1831 :     is now also part of the \pkg{surveillance} package.
1832 :     \item \code{polyCub.SV()} now also accepts objects of classes
1833 :     \code{"Polygon"} and \code{"Polygons"} for convenience.
1834 :     \item \code{siaf.lomax} is deprecated and replaced by
1835 :     \code{siaf.powerlaw} (re-parametrization).
1836 :     }
1837 :     }
1838 :    
1839 :     \subsection{NEW FEATURES (\code{twinstim()}-related)}{
1840 :     \itemize{
1841 :     \item The temporal \code{plot}-method for class \code{"epidataCS"}
1842 :     now understands the \code{add} parameter to add the histogram to an
1843 :     existing plot window, and auto-transforms the \code{t0.Date}
1844 :     argument using \code{as.Date()} if necessary.
1845 :     \item \code{nobs()} methods for classes \code{"epidataCS"} and
1846 :     \code{"twinstim"}.
1847 :     \item New argument \code{verbose} for \code{twinstim()} which, if
1848 :     set to \code{FALSE}, disables the printing of information messages
1849 :     during execution.
1850 :     \item New argument \code{start} for \code{twinstim()}, where (some)
1851 :     initial parameter values may be provided, which overwrite those in
1852 :     \code{optim.args$par}, which is no longer required (as a naive
1853 :     default, a crude estimate for the endemic intercept and zeroes for
1854 :     the other parameters are used).
1855 :     \item Implemented a wrapper \code{stepComponent()} for \code{step()}
1856 :     to perform algorithmic component-specific model selection in
1857 :     \code{"twinstim"} models. This also required the implementation of
1858 :     suitable \code{terms()} and \code{extractAIC()} methods. The single-step
1859 :     methods \code{add1()} and \code{drop1()} are also available.
1860 :     \item The \code{update.twinstim()} method now by default uses the
1861 :     parameter estimates from the previous model as initial values for
1862 :     the new fit (new argument \code{use.estimates = TRUE}).
1863 :     \item \code{as.epidataCS()} checks for consistency of the area of
1864 :     \code{W} and the (now really obligatory) area column in
1865 :     \code{stgrid}.
1866 :     \item \code{simulate.twinstim()} now by default uses the previous
1867 :     \code{nCircle2Poly} from the \code{data} argument.
1868 :     \item \code{direction} argument for \code{untie.epidataCS()}.
1869 :     \item The \code{toLatex}-method for \code{"summary.twinstim"} got
1870 :     different defaults and a new argument \code{eps.Pvalue}.
1871 :     \item New \code{xtable}-method for \code{"summary.twinstim"} for
1872 :     printing the covariate effects as risk ratios (with CI's and p-values).
1873 :     }
1874 :     }
1875 :    
1876 :     \subsection{NEW FEATURES (\code{hhh4()}-related)}{
1877 :     \itemize{
1878 :     \item New argument \code{hide0s} in the \code{plot}-method for class
1879 :     \code{"ah4"}.
1880 :     \item New argument \code{timevar} for \code{addSeason2formula()},
1881 :     which now also works for long formulae.
1882 :     }
1883 :     }
1884 :    
1885 :     }
1886 :    
1887 :    
1888 :    
1889 :     \section{Changes in surveillance version 1.5-1 (2012-12-14)}{
1890 :    
1891 :     \subsection{SYNOPSIS}{
1892 :     \itemize{
1893 :     \item The \pkg{surveillance} package is again backward-compatible
1894 :     with \R version 2.14.0, which is now declared as the minimum
1895 :     required version.
1896 :     }
1897 :     }
1898 :    
1899 :     }
1900 :    
1901 :    
1902 :    
1903 :     \section{Changes in surveillance version 1.5-0 (2012-12-12)}{
1904 :    
1905 :     \subsection{SYNOPSIS}{
1906 :     \itemize{
1907 :     \item This new version mainly improves upon the \code{twinstim()} and
1908 :     \code{hhh4()} implementations (see below).
1909 :     \item As requested by the CRAN team, examples now run faster. Some
1910 :     are conditioned on the value of the new package option
1911 :     \code{"allExamples"}, which usually defaults to \code{TRUE} (but is
1912 :     set to \code{FALSE} for CRAN checking, if timings are active).
1913 :     \item Moved some rarely used package dependencies to
1914 :     \dQuote{Suggests:}, and also removed some unused packages from there.
1915 :     \item Dropped strict dependence on \CRANpkg{gpclib}, which has a
1916 :     restricted license, for the \pkg{surveillance} package to be clearly
1917 :     GPL-2. Generation of \code{"epidataCS"} objects, which makes use of
1918 :     \pkg{gpclib}'s polygon intersection capabilities, now requires prior
1919 :     explicit acceptance of the \pkg{gpclib} license via setting
1920 :     \code{surveillance.options(gpclib = TRUE)}. Otherwise,
1921 :     \code{as.epidataCS()} and \code{simEpidataCS()} may not be used.
1922 :     }
1923 :     }
1924 :    
1925 :     \subsection{NEW FEATURES (\code{twinstim()}-related)}{
1926 :     \itemize{
1927 :     \item Speed-up by memoisation of the \code{siaf} cubature (using
1928 :     the \CRANpkg{memoise} package).
1929 :     \item Allow for \code{nlm}-optimizer (really not recommended).
1930 :     \item Allow for \code{nlminb}-specific control arguments.
1931 :     \item Use of the expected Fisher information matrix can be disabled
1932 :     for \code{nlminb} optimization.
1933 :     \item Use of the \code{effRange}-trick can be disabled in
1934 :     \code{siaf.gaussian()} and \code{siaf.lomax()}. The default
1935 :     \code{effRangeProb} argument for the latter has been changed from
1936 :     0.99 to 0.999.
1937 :     \item The \code{twinstim()} argument \code{nCub} has been replaced
1938 :     by the new \code{control.siaf} argument list. The old
1939 :     \code{nCub.adaptive} indicator became a feature of the
1940 :     \code{siaf.gaussian()} generator (named \code{F.adaptive} there) and
1941 :     does no longer depend on the \code{effRange} specification, but uses
1942 :     the bandwidth \code{adapt*sd}, where the \code{adapt} parameter may be
1943 :     specified in the \code{control.siaf} list in the \code{twinstim()}
1944 :     call. Accordingly, the components \code{"nCub"} and
1945 :     \code{"nCub.adaptive"} have been removed from the result
1946 :     of \code{twinstim()}, and are replaced by \code{"control.siaf"}.
1947 :     \item The \code{"method"} component of the \code{twinstim()} result
1948 :     has been replaced by the whole \code{"optim.args"}.
1949 :     \item The new \code{"Deriv"} component of \code{siaf} specifications
1950 :     integrates the \dQuote{siaf$deriv} function over a polygonal domain.
1951 :     \code{siaf.gaussian()} and \code{siaf.lomax()} use \code{polyCub.SV()}
1952 :     (with intelligent \code{alpha} parameters) for this task
1953 :     (previously: midpoint-rule with naive bandwidth)
1954 :     \item \code{scaled} \code{iafplot()} (default \code{FALSE}). The
1955 :     \code{ngrid} parameter has been renamed to \code{xgrid} and is more
1956 :     general.
1957 :     \item The \code{"simulate"} component of \code{siaf}'s takes an
1958 :     argument \code{ub} (upperbound for distance from the source).
1959 :     \item Numerical integration of spatial interaction functions with an
1960 :     \code{Fcircle} trick is more precise now; this slightly changes
1961 :     previous results.
1962 :     \item New \acronym{S3}-generic \code{untie()} with a method for the
1963 :     \code{"epidataCS"} class (to randomly break tied event times and/or
1964 :     locations).
1965 :     \item Renamed \code{N} argument of \code{polyCub.SV()} to
1966 :     \code{nGQ}, and \code{a} to \code{alpha}, which both have new
1967 :     default values.
1968 :     The optional polygon rotation proposed by Sommariva &
1969 :     Vianello is now also implemented (based on the corresponding MATLAB
1970 :     code) and available as the new \code{rotation} argument.
1971 :     \item The \code{scale.poly()} method for \code{"gpc.poly"} is now
1972 :     available as \code{scale.gpc.poly()}. The default return class of
1973 :     \code{discpoly()} was changed from \code{"gpc.poly"} to
1974 :     \code{"Polygon"}.
1975 :     \item An \code{intensityplot()}-method is now also implemented for
1976 :     \code{"simEpidataCS"}.
1977 :     }
1978 :     }
1979 :    
1980 :     \subsection{NEW FEATURES (\code{hhh4()}-related)}{
1981 :     \itemize{
1982 :     \item Significant speed-up (runs about 6 times faster now, amongst
1983 :     others by many code optimizations and by using sparse \CRANpkg{Matrix}
1984 :     operations).
1985 :     \item \code{hhh4()} optimization routines can now be customized for
1986 :     the updates of regression and variance parameters seperately, which
1987 :     for instance enables the use of Nelder-Mead for the variance
1988 :     updates, which seems to be more stable/robust as it does
1989 :     not depend on the inverse Fisher info and is usually faster.
1990 :     \item The \code{ranef()} extraction function for \code{"ah4"} objects
1991 :     gained a useful \code{tomatrix} argument, which re-arranges random
1992 :     effects in a unit x effect matrix (also transforming CAR effects
1993 :     appropriately).
1994 :     \item Generalized \code{hhh4()} to also capture parametric
1995 :     neighbourhood weights (like a power-law decay). The new function
1996 :     \code{nbOrder()} determines the neighbourhood order matrix
1997 :     from a binary adjacency matrix (depends on package \CRANpkg{spdep}).
1998 :     \item New argument \code{check.analyticals} (default \code{FALSE})
1999 :     mainly for development purposes.
2000 :     }
2001 :     }
2002 :    
2003 :     \subsection{BUG FIXES}{
2004 :     \itemize{
2005 :     \item Fixed sign of observed Fisher information matrix in
2006 :     \code{twinstim}.
2007 :     \item Simulation from the Lomax kernel is now correct (via polar
2008 :     coordinates).
2009 :     \item Fixed wrong Fisher information entry for the overdispersion
2010 :     parameter in \code{hhh4}-models.
2011 :     \item Fixed wrong entries in penalized Fisher information wrt the
2012 :     combination fixed effects x CAR intercept.
2013 :     \item Fixed indexing bug in penalized Fisher calculation in the case
2014 :     of multiple overdispersion parameters and random intercepts.
2015 :     \item Fixed bug in Fisher matrix calculation concerning the relation
2016 :     of unit-specific and random effects (did not work previously).
2017 :     \item Improved handling of non-convergent / degenerate solutions during
2018 :     \code{hhh4} optimization. This involves using \code{ginv()} from
2019 :     package \CRANpkg{MASS}, if the penalized Fisher info is singular.
2020 :     \item Correct labeling of overdispersion parameter in
2021 :     \code{"ah4"}-objects.
2022 :     \item Some control arguments of \code{hhh4()} have more clear
2023 :     defaults.
2024 :     \item The result of \code{algo.farrington.fitGLM.fast()} now
2025 :     additionally inherits from the \code{"lm"} class to avoid warnings
2026 :     from \code{predict.lm()} about fake object.
2027 :     \item Improved \file{NAMESPACE} imports.
2028 :     \item Some additional tiny bug fixes, see the subversion log on
2029 :     R-Forge for details.
2030 :     }
2031 :     }
2032 :    
2033 :     }
2034 :    
2035 :    
2036 :    
2037 :     \section{Changes in surveillance version 1.4-2 (2012-08-17)}{
2038 :    
2039 :     \subsection{SYNOPSIS}{
2040 :     \itemize{
2041 :     \item This is mainly a patch release for the \code{twinstim}-related
2042 :     functionality of the package.
2043 :     \item Apart from that, the package is now again compatible with older
2044 :     releases of \R (< 2.15.0) as intended (by defining \code{paste0()} in
2045 :     the package namespace if it is not found in \R \pkg{base} at
2046 :     installation of the \pkg{surveillance} package).
2047 :     }
2048 :     }
2049 :    
2050 :     \subsection{NEW FEATURES}{
2051 :     \itemize{
2052 :     \item Important new \code{twinstim()}-feature: fix parameters
2053 :     during optimization.
2054 :     \item Useful \code{update}-method for \code{"twinstim"}-objects.
2055 :     \item New \code{[[}- and \code{plot}-methods for
2056 :     \code{"simEpidataCSlist"}-objects.
2057 :     \item \code{simEpidataCS()} received tiny bug fixes and is now
2058 :     able to simulate from epidemic-only models.
2059 :     \item \code{R0}-method for \code{"simEpidataCS"}-objects (actually
2060 :     a wrapper for \code{R0.twinstim()}).
2061 :     \item Removed \code{dimyx} and \code{eps} arguments from
2062 :     \code{R0.twinstim()}; now uses \code{nCub} and
2063 :     \code{nCub.adaptive} from the fitted model and applies the same
2064 :     (numerical) integration method.
2065 :     \item \code{animate.epidata} is now compatible with the
2066 :     \CRANpkg{animation} package.
2067 :     \item More thorough documentation of \code{"twinstim"}-related
2068 :     functions \emph{including many examples}.
2069 :     }
2070 :     }
2071 :    
2072 :     \subsection{BUG FIXES (\code{"twinstim"}-related)}{
2073 :     \itemize{
2074 :     \item \code{nlminb} (instead of \code{optim}'s \code{"BFGS"}) is
2075 :     now the default optimizer (as already documented).
2076 :     \item The \code{twinstim}-argument \code{nCub} can now be omitted when
2077 :     using \code{siaf.constant()} (as documented) and is internally set to
2078 :     \code{NA_real_} in this case. Furthermore, \code{nCub} and
2079 :     \code{nCub.adaptive} are set to \code{NULL} if there is
2080 :     no epidemic component in the model.
2081 :     \item \code{toLatex.summary.twinstim} now again works for
2082 :     \code{summary(*, test.iaf=FALSE)}.
2083 :     \item \code{print}- and \code{summary}-methods for
2084 :     \code{"epidataCS"} no longer assume that the \code{BLOCK} index
2085 :     starts at 1, which may not be the case when using a subset in
2086 :     \code{simulate.twinstim()}.
2087 :     \item The \code{"counter"} step function returned by
2088 :     \code{summary.epidataCS()} does no longer produce false
2089 :     numbers of infectives (they were lagged by one timepoint).
2090 :     \item \code{plot.epidataCS()} now resolves \dots correctly and
2091 :     the argument \code{colTypes} takes care of a possible
2092 :     \code{subset}.
2093 :     \item \code{simEpidataCS()} now also works for endemic-only models
2094 :     and is synchronised with \code{twinstim()} regarding the
2095 :     way how \code{siaf} is numerically integrated (including the
2096 :     argument \code{nCub.adaptive}).
2097 :     \item Fixed problem with \code{simEpidataCS()} related to missing
2098 :     \file{NAMESPACE} imports (and re-exports) of \code{marks.ppp} and
2099 :     \code{markformat.default} from \CRANpkg{spatstat}, which are required
2100 :     for \code{spatstat::runifpoint()} to work, probably because
2101 :     \pkg{spatstat} currently does not register its S3-methods.
2102 :     \item Improved error handling in \code{simEpidataCS()}. Removed a
2103 :     \code{browser()}-call and avoid potentially infinite loop.
2104 :     }
2105 :     }
2106 :    
2107 :     \subsection{BUG FIXES (\code{"twinSIR"}-related)}{
2108 :     \itemize{
2109 :     \item The \code{.allocate} argument of \code{simEpidata()} has
2110 :     now a fail-save default.
2111 :     \item Simulation without endemic \code{cox()}-terms now works.
2112 :     }
2113 :     }
2114 :    
2115 :     \subsection{MINOR CHANGES}{
2116 :     \itemize{
2117 :     \item Simplified \code{imdepi} data to monthly instead of weekly
2118 :     intervals in \code{stgrid} for faster examples and reduced package
2119 :     size.
2120 :     \item The environment of all predefined interaction functions for
2121 :     \code{twinstim()} is now set to the \code{.GlobalEnv}. The previous
2122 :     behaviour of defining them in the \code{parent.frame()} could have
2123 :     led to huge \code{save()}'s of \code{"twinstim"} objects even with
2124 :     \code{model=FALSE}.
2125 :     \item \code{simulate.twinSIR} only returns a list of epidemics if
2126 :     \code{nsim > 1}.
2127 :     \item \code{simulate.twinstim} uses \code{nCub} and
2128 :     \code{nCub.adaptive} from fitted object as defaults.
2129 :     \item Removed the \dots-argument from \code{simEpidataCS()}.
2130 :     \item The coefficients returned by \code{simEpidataCS()} are now stored
2131 :     in a vector rather than a list for compatibility with
2132 :     \code{"twinstim"}-methods.
2133 :     \item Argument \code{cex.fun} of \code{intensityplot.twinstim()} now
2134 :     defaults to the \code{sqrt} function (as in \code{plot.epidataCS()}.
2135 :     }
2136 :     }
2137 :     }
2138 :    
2139 :    
2140 :     \section{Changes in surveillance version 1.4 (2012-07-26)}{
2141 :    
2142 :     \subsection{SYNOPSIS}{
2143 :     \itemize{
2144 :     \item Besides minor bug fixes, additional functionality has entered the package
2145 :     and a new attempt is made to finally release a new version on CRAN
2146 :     (version 1.3 has not appeared on CRAN), including a proper \file{NAMESPACE}.
2147 :     }
2148 :     }
2149 :    
2150 :     \subsection{NEW FEATURES}{
2151 :     \itemize{
2152 :     \item Support for non-parametric back-projection using the function
2153 :     \code{backprojNP()} which returns an object of the new
2154 :     \code{"stsBP"} class which inherits from \code{"sts"}.
2155 :     \item Bayesian nowcasting for discrete time count data is implemented in
2156 :     the function \code{nowcast()}.
2157 :     \item Methods for cubature over polygonal domains can now also visualize what
2158 :     they do. There is also a new quasi-exact method for cubature of the
2159 :     bivariate normal density over polygonal domains. The
2160 :     function \code{polyCub()} is a wrapper for the different
2161 :     methods.
2162 :     \item \code{residuals.twinstim()} and \code{residuals.twinSIR()}:
2163 :     extract the \dQuote{residual process}, see \cite{Ogata
2164 :     (1988)}. The residuals of \code{"twinSIR"} and
2165 :     \code{"twinstim"} models may be checked graphically by the new
2166 :     function \code{checkResidualProcess()}.
2167 :     \item Many new features for the \code{"twinstim"} class of
2168 :     self-exciting spatio-temporal point process models (see
2169 :     below).
2170 :     }
2171 :     }
2172 :    
2173 :     \subsection{NEW FEATURES AND SIGNIFICANT CHANGES FOR \code{"twinstim"}}{
2174 :     \itemize{
2175 :     \item Modified arguments of \code{twinstim()}: new ordering, new
2176 :     argument \code{nCub.adaptive}, removed argument
2177 :     \code{typeSpecificEndemicIntercept} (which is now specified as part of
2178 :     the \code{endemic} formula as \code{(1|type)}).
2179 :     \item Completely rewrote the \code{R0}-method (calculate \dQuote{trimmed} and
2180 :     \dQuote{untrimmed} \eqn{R_0} values)
2181 :     \item The \dQuote{trimmed} \code{R0} values are now part of the
2182 :     result of the model fit, as well as \code{bbox(W)}. The
2183 :     model evaluation environment is now set as attribute of the
2184 :     result if \code{model=TRUE}.
2185 :     \item New predefined spatial kernel: the Lomax power law kernel
2186 :     \code{siaf.lomax()}
2187 :     \item \code{plot}-methods for \code{"twinstim"}
2188 :     (\code{intensityplot()} and \code{iafplot()})
2189 :     \item \code{as.epidataCS()} now auto-generates the stop-column if this is missing
2190 :     \item \code{print}-method for class \code{"summary.epidataCS"}
2191 :     \item \code{[}- and subset-method for \code{"epidataCS"}
2192 :     (subsetting \code{...$events})
2193 :     \item \code{plot}-method for \code{"epidataCS"}
2194 :     }
2195 :     }
2196 :    
2197 :     \subsection{MINOR CHANGES}{
2198 :     \itemize{
2199 :     \item Improved documentation for the new functionalities.
2200 :     \item Updated references.
2201 :     \item \code{twinSIR}'s \code{intensityPlot} is now a method of the
2202 :     new S3-generic function \code{intensityplot}.
2203 :     }
2204 :     }
2205 :     }
2206 :    
2207 :    
2208 :     \section{Changes in surveillance version 1.3 (2011-04-25)}{
2209 :     \subsection{SYNOPSIS}{
2210 :     \itemize{
2211 :     \item This is a major realease integrating plenty of new code (unfortunately
2212 :     not all documented as good as it could be). This includes code
2213 :     for the \code{"twinstim"} and the \code{"hhh4"} model.
2214 :     The \code{"twinSIR"} class of models has been
2215 :     migrated from package \pkg{RLadyBug} to \pkg{surveillance}.
2216 :     It may take a while before this version will become available from CRAN.
2217 :     For further details see below.
2218 :     }
2219 :     }
2220 :     \subsection{SIGNIFICANT CHANGES}{
2221 :     \itemize{
2222 :     \item Renamed the \code{"week"} slot of the \code{"sts"} S4 class to \code{"epoch"}.
2223 :     All saved data objects have accordingly be renamed, but some hazzle
2224 :     is to be expected if one you have old \code{"sts"} objects stored in binary
2225 :     form. The function \code{convertSTS()} can be used to
2226 :     convert such \dQuote{old school} \code{"sts"} objects.
2227 :     \item Removed the functions \code{algo.cdc()} and \code{algo.rki()}.
2228 :     }
2229 :     }
2230 :     \subsection{NEW FEATURES}{
2231 :     \itemize{
2232 :     \item Support for \code{"twinSIR"} models (with associated
2233 :     \code{"epidata"} objects) as described
2234 :     in \enc{Höhle}{Hoehle} (2009) has been moved from package
2235 :     \pkg{RLadyBug} to \pkg{surveillance}.
2236 :     That means continuous-time discrete-space \acronym{SIR} models.
2237 :     \item Support for \code{"twinstim"} models as described in
2238 :     \cite{Meyer et al (2012)}. That means continuous-time
2239 :     continuous-space infectious disease models.
2240 :     \item Added functionality for non-parametric back projection
2241 :     (\code{backprojNP()}) and
2242 :     now-casting (\code{nowcast()}) based on \code{"sts"} objects.
2243 :     }
2244 :     }
2245 :     }
2246 :    
2247 :     \section{Changes in surveillance version 1.2-2}{
2248 :     \itemize{
2249 :     \item Replaced the deprecated getSpPPolygonsLabptSlots method with calls
2250 :     to the coordinates method when plotting the map slot.
2251 :     \item Minor proof-reading of the documentation.
2252 :     \item Added an argument \code{"extraMSMargs"} to the algo.hmm function.
2253 :     \item Fixed bug in \code{outbreakP()} when having observations equal to zero
2254 :     in the beginning. Here, \eqn{\hat{\mu}^{C1}} in (5) of \cite{Frisen et al (2008)}
2255 :     is zero and hence the log-based summation in the code failed.
2256 :     Changed to product as in the original code, which however might be
2257 :     less numerically stable.
2258 :     \item Fixed bug in stcd which added one to the calculated index of idxFA and idxCC.
2259 :     Thanks to Thais Rotsen Correa for pointing this out.
2260 :     }
2261 :     }
2262 :    
2263 :     \section{Changes in surveillance version 1.2-1 (2010-06-10)}{
2264 :     \itemize{
2265 :     \item Added \code{algo.outbreakP()} (\cite{Frisen & Andersson, 2009}) providing a
2266 :     semiparametric approach for outbreak detection for Poisson
2267 :     distributed variables.
2268 :     \item Added a pure \R function for extracting ISO week and year from Date
2269 :     objects. This function (isoWeekYear) is only called if "\%G" and "\%V"
2270 :     format strings are used on Windows (\code{sessionInfo()[[1]]$os == "mingw32"})
2271 :     as this is not implemented for \code{"format.Date"} on Windows.
2272 :     Thanks to Ashley Ford, University of Warwick, UK for identifying
2273 :     this Windows specific bug.
2274 :     \item For \code{algo.farrington()} a faster fit routine \code{"algo.farrington.fitGLM.fast"}
2275 :     has been provided by Mikko Virtanen, National Institute for Health
2276 :     and Welfare, Finland. The new function calls \code{glm.fit()}
2277 :     directly, which gives a doubling of speed for long series. However, if one
2278 :     wants to process the fitted model output some of the GLM routines might
2279 :     not work on this output. For backwards compability the argument
2280 :     \code{control$fitFun = "algo.farrington.fitGLM"} provides the old (and slow)
2281 :     behaviour.
2282 :     }
2283 :     }
2284 :    
2285 :     \section{Changes in surveillance version 1.1-6 (2010-05-25)}{
2286 :     \itemize{
2287 :     \item A few minor bug fixes
2288 :     \item Small improvements in the C-implementation of the \code{twins()}
2289 :     function by Daniel Saban\enc{é}{e}s Bov\enc{é}{e} fixing the segmentation fault
2290 :     issue on 64-bit architectures.
2291 :     }
2292 :     }
2293 :    
2294 :     \section{Changes in surveillance version 1.1-2 (2009-10-15)}{
2295 :     \itemize{
2296 :     \item Added the functions categoricalCUSUM and LRCUSUM.runlength
2297 :     for the CUSUM monitoring of general categorical time series
2298 :     (binomial, beta-binomial, multinomial, ordered response,
2299 :     Bradley-Terry models).
2300 :     \item Added the functions pairedbinCUSUM and pairedbinCUSUM.runlength
2301 :     implementing the CUSUM monitoring and run-length computations for
2302 :     a paired binary outcome as described in Steiner et al. (1999).
2303 :     \item Experimental implementation of the prospective space-time cluster
2304 :     detection described in Assuncao and Correa (2009).
2305 :     \item Added a \code{demo("biosurvbook")} containing the code of an upcoming
2306 :     book chapter on how to use the surveillance package. This
2307 :     contains the description of ISO date use, negative binomial CUSUM,
2308 :     run-length computation, etc. From an applicational point of view
2309 :     the methods are illustrated by Danish mortality monitoring.
2310 :     \item Fixed a small bug in algo.cdc found by Marian Talbert Allen
2311 :     which resulted in the control$m argument being ignored.
2312 :     \item The constructor of the sts class now uses the argument
2313 :     \code{"epoch"} instead of weeks to make clearer that also
2314 :     daily, monthly or other data can be handled.
2315 :     \item Added additional epochAsDate slot to sts class. Modified
2316 :     plot functions so they can handle ISO weeks.
2317 :     \item algo.farrington now also computes quantile and median of
2318 :     the predictive distribution. Furthermore has the computation
2319 :     of reference values been modified so its a) a little bit faster
2320 :     and b) it is also able to handle ISO weeks now. The reference values
2321 :     for date t0 are calculated as follows:
2322 :     For i, i=1,..., b look at date t0 - i*year. From this date on move
2323 :     w months/weeks/days to the left and right. In case of weeks:
2324 :     For each of these
2325 :     determined time points go back in time to the closest Monday
2326 :     \item Renamed the functions obsinyear to epochInYear, which now also
2327 :     handles objects of class Date.
2328 :     }
2329 :     }
2330 :    
2331 :     \section{Changes in surveillance version 1.0-2 (2009-03-06)}{
2332 :     \itemize{
2333 :     \item Negative Binomial CUSUM or the more general NegBin likelihood ratio
2334 :     detector is now implemented as part of algo.glrnb.
2335 :     This includes the back calculation of the required number of cases
2336 :     before an alarm.
2337 :     \item Time varying proportion binomial CUSUM.
2338 :     }
2339 :     }
2340 :    
2341 :     \section{Changes in surveillance version 0.9-10}{
2342 :     \itemize{
2343 :     \item Current status: Development version available from
2344 :     \url{http://surveillance.r-forge.r-project.org/}
2345 :     \item Rewriting of the plot.sts.time.one function to use polygons
2346 :     instead of lines for the number of observed cases. Due cause
2347 :     a number of problems were fixed in the plotting of the legend.
2348 :     Plotting routine now also handles binomial data, where the
2349 :     number of observed cases y are stored in \code{"observed"} and the
2350 :     denominator data n are stored in \code{"populationFrac"}.
2351 :     \item Problems with the aggregate function not operating correctly
2352 :     for the populationFrac were fixed.
2353 :     \item The \code{"rogerson"} wrapper function for algo.rogerson was modified so it
2354 :     now works better for distribution \code{"binomial"}. Thus a time varying
2355 :     binomial cusum can be run by calling
2356 :     \code{rogerson( x, control(..., distribution="binomial"))}
2357 :     \item An experimental implementation of the twins model documented in
2358 :     Held, L., Hofmann, M., \enc{Höhle}{Hoehle}, M. and Schmid V. (2006). A two-component
2359 :     model for counts of infectious diseases, Biostatistics, 7, pp.
2360 :     422--437 is now available as algo.twins.
2361 :     }
2362 :     }
2363 :    
2364 :     \section{Changes in surveillance version 0.9-9 (2008-01-21)}{
2365 :     \itemize{
2366 :     \item Fixed a few small problems
2367 :     which gave warnings in the CRAN distribution
2368 :     }
2369 :     }
2370 :    
2371 :     \section{Changes in surveillance version 0.9-8 (2008-01-19)}{
2372 :     \itemize{
2373 :     \item The algo_glrpois function now has an additional \code{"ret"} arguments,
2374 :     where one specifies the return type. The arguments of the underlying
2375 :     c functions have been changed to include an additional direction and
2376 :     return type value arguments.
2377 :     \item added restart argument to the algo.glrpois control object, which
2378 :     allows the user to control what happens after the first alarm has been
2379 :     generated
2380 :     \item experimental algo.glrnb function is added to the package. All calls to
2381 :     algo.glrpois are now just alpha=0 calls to this function. However,
2382 :     the underlying C functions differentiate between poisson and negative case
2383 :     }
2384 :     }

R-Forge@R-project.org
ViewVC Help
Powered by ViewVC 1.0.0  
Thanks to:
Vienna University of Economics and Business University of Wisconsin - Madison Powered By FusionForge