RELEASE NOTES

spatstat 1.44-0

22 December 2015

We thank Jonas Geldmann, Andrew Hardegen, Kassel Hingee, Tom Lawrence, Robin Milne, Gopalan Nair, Suman Rakshit, Peijian Shi and Rasmus Waagepetersen for contributions.

OVERVIEW

More support for multidimensional point patterns and point processes.
More options for envelopes and related Monte Carlo tests.
More support for model comparison.
k-th nearest neighbours on a linear network.
Penttinen process can be simulated (by Metropolis-Hastings or CFTP).
Calculate the predicted variance of number of points.
Convexifying operation for sets.
Subdivide a linear network.
Accelerated algorithms for linear networks.
Quadrat counting accelerated, in some cases.
Version nickname: “The Sound of One Hand Typing”

NEW FUNCTIONS

rPenttinen Simulate the Penttinen process using perfect simulation.
varcount Given a point process model, compute the predicted variance of the number of points falling in a window.
inside.boxx Test whether multidimensional points lie inside a specified multidimensional box.
lixellate Divide each segment of a linear network into smaller segments.
nsegments.linnet,nsegments.lpp Count the number of line segments in a linear network.
grow.boxx Expand a multidimensional box.
deviance.ppm,deviance.lppm Deviance for a fitted point process model.
pseudoR2 Pseudo-R-squared for a fitted point process model.
tiles.empty Checks whether each tile of a tessellation is empty or nonempty.
summary.linim Summary for a pixel image on a linear network.

SIGNIFICANT USER-VISIBLE CHANGES

rMaternI,rMaternII These functions can now generate random patterns in three dimensions and higher dimensions, when the argument “win” is of class “box3” or “boxx”.
”[.ppx” The subset index ‘i’ may now be a spatial domain of class ‘boxx’ or ‘box3’.
rmh.ppm, rmhmodel.ppm, simulate.ppm A model fitted using the ‘Penttinen’ interaction can now be simulated.
rmh.default, rmhmodel.default These functions now recognise cif=’penttinen’ for the Penttinen interaction.
envelope New argument ‘clamp’ gives greater control over one-sided envelopes.
dclf.test, mad.test, dclf.progress, mad.progress, dclf.sigtrace, mad.sigtrace New argument ‘clamp’ determines the test statistic for one-sided tests.
dclf.progress, mad.progress, dclf.sigtrace, mad.sigtrace, mctest.progress, mctest.sigtrace, dg.progress, dg.sigtrace New argument ‘rmin’ determines the left endpoint of the test interval.
dclf.test, mad.test, dg.test, dg.progress, dg.sigtrace, dg.envelope (also accepted by dclf.progress, mad.progress, dclf.sigtrace, mad.sigtrace) New argument ‘leaveout’ specifies how to calculate the deviation between the observed summary function and nominal reference value.
envelope New argument ‘funargs’
Hest Argument X can now be a pixel image with logical values. New argument ‘W’. [Based on code by Kassel Hingee.]
nncross.lpp, distfun.lpp New argument ‘k’ allows calculation of k-th nearest neighbour. Computation accelerated.
logLik.ppm New argument ‘absolute’.
plot.kppm New arguments ‘pause’ and ‘xname’.
tess Argument ‘window’ is ignored when xgrid, ygrid are given.
as.polygonal Can now repair errors in polygon data, if repair=TRUE.
rStrauss, rHardcore, rStraussHard, rDiggleGratton, rDGS, rPenttinen New argument ‘drop’.
Kest.fft Now has ‘…’ arguments allowing control of spatial resolution.
lppm Computation accelerated.
quadratcount.ppp Computation accelerated in some cases.
dg.test Computation accelerated.

BUG FIXES

runifpointx, rpoisppx Crashed if nsim > 1. Fixed.
triangulate.owin Results were incorrect in some special cases. Fixed.
quadrat.test, clarkevans.test In rare cases, the computed Monte Carlo p-value could have been greater than 1. This could have occurred only when nsim was an even number and when the correct p-value was equal to 1. Fixed.
linearmarkequal Result was a data frame instead of an ‘fv’ object. Fixed.
point-in-polygon test The function inside.owin could take a very long time to check whether points are inside a polygonal window, if the coordinates were very large numbers. This was due to numerical overflow. (Fixed??)
as.fv.kppm Crashed if the model was not fitted by minimum contrast. Fixed.
plot.fv Crashed in some obscure cases. Fixed.
collapse.fv Did not allow ‘same=NULL’. Fixed.
dclf.progress, mad.progress, dg.progress, dclf.sigtrace, mad.sigtrace, dg.sigtrace The results could not be re-plotted using a plot formula, because the internal data were slightly corrupted. Fixed.
Kest.fft Result was incorrectly normalised. Fixed.
crosspairs If X and Y were identical point patterns, the result was not necessarily symmetric (on some machines) due to numerical artifacts. Fixed.
plot.fv Lines were not correctly clipped to the plot region when ‘ylim’ was given. Fixed.
pool.envelope The ‘scale’ argument was not handled correctly. Fixed.

Release notes are available in raw text format here.

Spatstat