RELEASE NOTES
spatstat 1.45-2
9 May 2016
We thank Ottmar Cronie, Virginia Morera Pujol, Sven Wagner and Marie-Colette van Lieshout for contributions.
OVERVIEW
-
Recursive-partition point process models.
-
Minkowski sum, morphological dilation and erosion with any shape.
-
Important bug fix in spatial CDF tests.
-
More bug fixes for replicated patterns.
-
Simulate a model fitted to replicated point patterns.
-
Inhomogeneous multitype F and G functions.
-
Summary functions recognise correction=”all”
-
Leverage and influence code handles bigger datasets.
-
More support for pixel images.
-
Improved progress reports.
-
New dataset ‘redwood3’
-
spatstat now Depends on the package ‘rpart’
-
Version nickname: ‘Caretaker Mode’
NEW FUNCTIONS
-
as.im.data.frame Build a pixel image from a data frame of coordinates and pixel values.
-
covering Cover a window using discs of a given radius.
-
dilationAny, erosionAny, %(-)% Morphological dilation and erosion by any shape.
-
FmultiInhom, GmultiInhom Inhomogeneous multitype/marked versions of the summary functions Fest, Gest.
-
kernel.moment Moment or incomplete moment of smoothing kernel.
-
MinkowskiSum, %(+)% Minkowski sum of two windows: A %(+)% B, or MinkowskiSum(A,B)
-
nobjects New generic function for counting the number of ‘things’ in a dataset. There are methods for ppp, ppx, psp, tess.
-
parameters.interact, parameters.fii Extract parameters from interpoint interactions. [These existing functions are now documented.]
-
ppmInfluence Calculate leverage.ppm, influence.ppm and dfbetas.ppm efficiently.
-
rppm, plot.rppm, predict.rppm, prune.rppm Recursive-partition point process models
-
simulate.mppm Simulate a point process model fitted to replicated point patterns.
-
update.interact Update the parameters of an interpoint interaction. [This existing function is now documented.]
-
where.max, where.min Find the spatial location(s) where a pixel image achieves its maximum or minimum value.
SIGNIFICANT USER-VISIBLE CHANGES
-
cdf.test.mppm Now handles Gibbs models. Now recognises covariate=”x” or “y”.
-
leverage.ppm, influence.ppm, dfbetas.ppm For Gibbs models, memory usage has been dramatically reduced, so the code can handle larger datasets and finer quadrature schemes.
-
plot.im Now handles complex-valued images.
-
connected.im Now handles a logical-valued image properly.
-
qqplot.ppm Argument ‘expr’ can now be a list of point patterns, or an envelope object containing a list of point patterns.
-
as.layered Default method now handles a (vanilla) list of spatial objects.
-
summary functions The argument ‘correction=”all”’ is now recognised: it selects all the available options. This applies to Fest, F3est, Gest, Gcross, Gdot, Gmulti, G3est, Gfox, Gcom, Gres, Hest, Jest, Jmulti, Jcross, Jdot, Jfox, Kest, Kinhom, Kmulti, Kcross, Kdot, Kcom, Kres, Kmulti.inhom, Kcross.inhom, Kdot.inhom, Kscaled, Ksector, Kmark, K3est, Lscaled, markcorr, markcrosscorr, nnorient, pairorient, pcfinhom, pcfcross.inhom, pcfcross, pcf, Tstat.
-
clarkevans The argument ‘correction=”all”’ is now recognised: it selects all the available options. [This is also the default.]
-
predict.mppm The argument ‘type=”all”’ is now recognised: it selects all the available options. [This is also the default.]
-
plot.kppm The argument ‘what=”all”’ is now recognised: it selects all the available options. [This is also the default.]
-
connected.im, connected.owin Arguments ‘…’ now determine pixel resolution.
-
anova.mppm New argument ‘fine’
-
as.owin.data.frame New argument ‘step’
-
discs Now accepts a single numeric value for ‘radii’.
-
plot.ppp, plot.profilepl, plot.quadratcount, plot.quadrattest, plot.tess Now recognise graphics parameters for text, such as ‘family’ and ‘srt’
-
as.function.tess New argument ‘values’ specifies the function values.
-
cdf.test Calculations are more robust against numerical rounding effects.
-
progressreport Behaviour improved. New arguments ‘tick’, ‘showtime’.
-
simulate.ppm New argument ‘verbose’
-
compileK, compilepcf These internal functions are now documented.
BUG FIXES
-
cdf.test.ppm Calculation of p-values was incorrect for Gibbs models: 1-p was computed instead of p. [Spotted by Sven Wagner.] Fixed.
-
subfits The interaction coefficients of the submodels were incorrect for Gibbs models with a multitype interaction (MultiStrauss, etc). [Spotted by Sven Wagner.] Fixed.
-
subfits Crashed when a Gibbs model included factor-valued spatial covariates and not all levels of the factor were present in each row of the data. [Spotted by Sven Wagner.] Fixed.
-
subfits For Gibbs models with a multitype interaction (MultiStrauss, etc), computation of the conditional intensity caused an error. [Spotted by Sven Wagner.] Fixed.
-
diagnose.ppm Crashed if what=”smooth”, when the original window was a rectangle. [Spotted by Virginia Morera Pujol.] Fixed.
-
mppm The x and y coordinates were not permitted in the random-effects formula ‘random’. [Spotted by Sven Wagner.] Fixed.
-
vcov.ppm The result had no ‘dimnames’, if the model was fitted using method=”ho”. Fixed.