Spatstat

analysing spatial point patterns

• Home
• News
• Download
• Docs
• Book
• Courses
• Help
• FAQ
• About

RELEASE NOTES

spatstat 1.43-0

07 October 2015

We thank Leanne Bischof, Christophe Biscio, Belarmain Fandohan, Andrew Hardegen, Frederic Lavancier, Tom Lawrence, Martin Maechler, Greg McSwiggan, Robin Milne, Gopalan Nair, Tuomas Rajala, Suman Rakshit, Ben Ramage, Francois Semecurbe and Ida-Maria Sintorn for contributions.

OVERVIEW

• spatstat now depends on the package ‘nlme’.

• spatstat now depends on R 3.2.2 or later.

• Simulation algorithms have been accelerated; simulation outcomes
  are not identical to those obtained from previous versions of spatstat.

• Determinantal point process models.

• Random-effects and mixed-effects models for replicated patterns.

• Dao-Genton test, and corresponding simulation envelopes.

• Simulated annealing and simulated tempering.

• spatstat colour tools now handle transparent colours.

• Improvements to “[” and subset() methods

• Extensions to kernel smoothing on a linear network.

• Support for one-dimensional smoothing kernels.

• Bug fix in Metropolis-Hastings simulation.

• Mark correlation function may include weights.

• Cross-correlation version of the mark correlation function.

• Variance calculations for replicated patterns.

• Penttinen pairwise interaction model.

• Contour plots with colours determined by a colour map.

• New dataset: Australian states and territories.

• More support for multi-dimensional point patterns.

• Minor improvements and bug fixes.

• Version nickname: “Mixed Effects”

NEW DATASET

• austates
  The states and large mainland territories of Australia
  represented as polygonal regions forming a tessellation.

NEW FUNCTIONS

• dppm
  Fit a determinantal point process model to point pattern data.

• fitted.dppm, predict.dppm, intensity.dppm
  Predict a fitted dppm object.

• logLik.dppm, AIC.dppm, extractAIC.dppm, nobs.dppm
  Likelihood and AIC for determinantal point process models
  (enabling the use of ‘step’)

• coef.dppm, formula.dppm, print.dppm, terms.dppm, labels.dppm,
  model.frame.dppm, model.matrix.dppm, model.images.dppm,
  is.stationary.dppm, reach.dppm, unitname.dppm, unitname<-.dppm, Window.dppm
  Various methods for dppm objects.

• parameters.dppm
  Extract meaningful list of model parameters

• objsurf.dppm
  Objective function surface of a dppm object

• residuals.dppm
  Residual measure for a dppm object.

• dppBessel, dppCauchy, dppGauss, dppMatern, dppPowerExp
  Determinantal Point Process models.

• update.dppmodel
  Set parameter values in a dpp model.

• is.stationary.dppmodel, print.dppmodel, reach.dppmodel, valid.dppmodel
  Basic information about a dpp model

• rdpp, simulate.dppmodel
  Simulation of a dpp model.

• intensity.dppmodel, Kmodel.dppmodel, pcfmodel.dppmodel
  Moments of a dpp model

• dim.dppmodel, dppapproxkernel, dppapproxpcf, dppeigen, dppfamily,
  dppkernel, dppparbounds, dppspecdenrange, dppspecden
  Helper functions for dpp models.

• dclf.sigtrace, mad.sigtrace, mctest.sigtrace
  Significance trace of Monte Carlo test

• dg.test
  Dao-Genton adjusted Monte Carlo goodness-of-fit test.

• dg.envelope
  Simulation envelopes corresponding to Dao-Genton test.

• dg.sigtrace
  Significance trace for Dao-Genton test

• dg.progress
  Progress plot for Dao-Genton test

• markcrosscorr
  Mark cross-correlation function for point patterns with
  several columns of marks

• fixef.mppm, ranef.mppm
  Extract fixed effects and random effects from a point process model
  fitted to replicated point patterns.

• rtemper
  Simulated annealing or simulated tempering.

• to.opaque, to.transparent
  Change transparency value in colours

• rgb2hsva
  Convert RGB to HSV data, like rgb2hsv, but preserving transparency.

• superimpose.ppplist, superimpose.splitppp
  New methods for ‘superimpose’ for lists of point patterns.

• dkernel, pkernel, qkernel, rkernel
  Probability density, cumulative probability, quantiles
  and random generation from distributions used in basic one-dimensional
  kernel smoothing.

• kernel.factor
  Auxiliary calculations for one-dimensional kernel smoothing.

• PPversion, QQversion
  Transformation of a summary function to its P-P or Q-Q counterpart.

• spatdim
  Spatial dimension of any object in the spatstat package.

• as.boxx
  Convert data to a multi-dimensional box.

• intensity.ppx
  Method for ‘intensity’ for multi-dimensional space-time point patterns.

• fourierbasis
  Evaluate Fourier basis functions in any number of dimensions.

• valid
  New generic function, with methods valid.ppm, valid.lppm, valid.dppmodel

• emend, emend.ppm, emend.lppm
  New generic function with methods for ppm and lppm.
  emend.ppm is equivalent to project.ppm

• Penttinen
  New pairwise interaction model.

• quantile.density
  Calculates quantiles from kernel density estimates.

• CDF.density
  Calculates cumulative distribution function from kernel density estimates.

SIGNIFICANT USER-VISIBLE CHANGES

• simulation
  Several basic simulation algorithms have been accelerated.
  Consequently, simulation outcomes are not identical to
  those obtained with previous versions of spatstat, even when the
  same random seed is used. To ensure compatibility with previous
  versions of spatstat, revert to the slower code by setting
  spatstat.options(fastthin=FALSE, fastpois=FALSE).

• mppm
  Now handles models with a random effect component.
  New argument ‘random’ is a formula specifying the random effect.

• vcov.mppm
  Now handles models with Gibbs interactions.

• [.ppp
  New argument ‘clip’ determines whether the window is clipped.

• [.ppp
  The previously-unused argument ‘drop’ now determines whether
  to remove unused levels of a factor.

• [.pp3, [.lpp, [.ppx, subset.ppp, subset.pp3, subset.lpp, subset.ppx
  These methods now have an argument ‘drop’ which determines
  whether to remove unused levels of a factor.

• density.lpp
  Now supports both the ‘equal-split continuous’ and
  ‘equal-split discontinuous’ smoothers. New argument ‘continuous’
  determines the choice of smoother.

• envelope
  New argument ‘scale’ allows global envelopes to have
  width proportional to a specified function of r,
  rather than constant width.

• dclf.test, mad.test, dclf.progress, mad.progress, mctest.progress
  New argument ‘scale’ allows summary function values to be rescaled
  before the comparison is performed.

• dclf.test, mad.test
  New argument ‘interpolate’ supports interpolation of p-value.

• dclf.progress, mad.progress, mctest.progress
  New argument ‘interpolate’ supports interpolation of critical value of test.

• simulate.ppm
  New argument ‘w’ controls the window of the simulated patterns.

• default.rmhcontrol, default.rmhexpand
  New argument ‘w’.

• markcorr
  New argument ‘weights’ allows computation of the weighted version
  of the mark correlation function.

• density.lpp
  New argument ‘kernel’ specifies the smoothing kernel.
  Any of the standard one-dimensional smoothing kernels can be used.

• contour.im
  New argument ‘col’ specifies the colour of the contour lines.
  If ‘col’ is a colour map, then the contours are drawn in different colours.

• plot.ppp
  The default colour for the points is now a transparent grey,
  if this is supported by the plot device.

• rgbim, hsvim
  New argument ‘A’ controls the alpha (transparency) channel.

• rgb2hex, col2hex, paletteindex, is.colour, samecolour,
  complementarycolour, is.grey, to.grey
  These colour tools now handle transparent colours.

• rgb2hex
  New argument ‘maxColorValue’

• to.grey
  New argument ‘transparent’.

• progressreport
  New argument ‘state’
  New option: style=”tk”

• rLGCP
  This function no longer requires the package ‘RandomFields’
  to be loaded explicitly.

• kppm
  Fitting a model with clusters=”LGCP” no longer requires the
  package ‘RandomFields’ to be loaded explicitly.

• rpoispp
  Accelerated, when ‘lambda’ is a pixel image.

• rthin
  Accelerated, when ‘P’ is a single number.

• spatstat.options
  New options ‘fastthin’ and ‘fastpois’ enable fast simulation algorithms.
  Set these options to FALSE to reproduce results obtained with
  previous versions of spatstat.

• split.ppp
  The splitting variable ‘f’ can now be a logical vector.

• collapse.fv
  This is now treated as a method for the ‘nlme’ generic ‘collapse’.
  Its syntax has been adjusted slightly.

• diagnose.ppm, plot.diagppm
  New arguments col.neg, col.smooth control the colour maps.

• valid.ppm
  This is now a method for the generic function ‘valid’.

• ppm.ppp, ppm.quad
  New argument ‘emend’, equivalent to ‘project’.

• ”[<-.im”
  Accepts an array for ‘value’.

• as.im.function
  New argument ‘strict’.

• bw.ppl
  New argument ‘weights’.

• plot.mppm
  New argument ‘se’.

• dclf.test, mad.test
  Formal arguments ‘use.theo’ and ‘internal’ have been removed.

• predict.kppm, residuals.kppm
  Now issues a warning when the calculation ignores the
  cluster/Cox component and treats the model as if it were Poisson.
  (This currently happens in predict.kppm when se=TRUE or interval != “none”,
  and in residuals.kppm when type != “raw”).

BUG FIXES

1. lpp
   Crashed if X was a 4-column matrix.
   Fixed.

2. plot.fv
   Crashed with some graphics devices, if legend=TRUE.
   Fixed.

3. effectfun
   Crashed if ‘covname’ was missing.
   Fixed.

4. rVarGamma, rMatClust, rThomas, rCauchy, rNeymanScott
   Crashed if ‘kappa’ was a function or image instead of a single number.
   [Spotted by Ben Ramage.]
   Fixed.

5. plot.mppm
   Crashed with a message about “figure margins too large”
   unless the argument se=FALSE was given explicitly.
   Fixed.

6. opening.owin, closing.owin
   Crashed sometimes, with a message about a rectangle not containing a window.
   Fixed.

7. persp.im
   Crashed if all pixel values were equal to zero (unless zlim was given).
   Fixed.

8. predict.ppm
   Crashed sometimes if the model was fitted with use.gam=TRUE.

9. as.linim.linfun
   Generated an error (‘L must be a linear network’)
   if extra arguments were given.

10. as.function.fv
    Generated an error when executed in the ‘covr’ package.
    Fixed.

11. rmh, simulate.ppm
    Results were incorrect for inhomogeneous multitype models
    simulated with fixall=TRUE (i.e. prescribing a fixed number of
    points of each type) if the model was segregated
    (i.e. if different types of points had different first order trend).
    Fixed.

12. dclf.progress, mad.progress
    Ignored the argument ‘alternative’.
    Fixed.

13. $<-.hyperframe, [<-.hyperframe
    Result was garbled if ‘value’ was a hyperframe with one column.

14. rmh.ppm
    Argument ‘w’ was ignored in some cases.
    Fixed.

15. Hest
    There was an artefact at r=0 when conditional=TRUE.
    Fixed.

16. [.msr
    The result of M[W] where W is a window
    was a measure with window W, instead of intersect.owin(W, Window(M)).
    Fixed.

17. pool.envelope
    Did not always respect the value of ‘use.theory’.
    Fixed.

18. envelope, pool.envelope
    If ‘ginterval’ was given, the results were in a slightly incorrect format.
    Fixed.

19. pool.envelope
    Did not check for compatible values of ‘ginterval’.
    Fixed.

Release notes are available in raw text format here.