Spatstat

analysing spatial point patterns

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

RELEASE NOTES

spatstat 1.61-0

12 September 2019

We thank Jordan Brown, Tilman Davies and Greg McSwiggan for contributions.

OVERVIEW

• Fast kernel estimation on a linear network using 2D kernels.

• Nonparametric maximum likelihood estimation of ‘rho’.

• Extension of Scott’s rule for bandwidth selection.

• Cross-validated bandwidth selection on a linear network.

• More support for character-valued images.

• Random thinning of clumps.

• Bug fixes and minor improvements.

• Nickname: ‘Puppy zoomies’

NEW FUNCTIONS

• densityQuick.lpp
  Fast kernel estimator of point process intensity on a network
  using 2D smoothing kernel.

• bw.scott.iso
  Isotropic version of Scott’s rule (for point patterns in any dimension).

• data.lppm
  Extract the original data point pattern from a fitted model of class ‘lppm’.

• rthinclumps
  Divide a spatial region into clumps and randomly delete some of them.

• dimnames.hyperframe, dimnames<-.hyperframe
  Methods for extracting and changing the ‘dimnames’ of a hyperframe.

SIGNIFICANT USER-VISIBLE CHANGES

• rhohat
  Estimation by nonparametric maximum likelihood is now supported,
  assuming the intensity is a monotone function of the covariate.
  New options: smoother=”increasing” and smoother=”decreasing”.

• density.lpp
  New argument ‘distance’ offers a choice of different kernel methods.

• bw.scott
  Now handles point patterns of any dimension.
  New arguments ‘isotropic’ and ‘d’.

• bw.ppl
  New argument ‘shortcut’ allows faster computation.
  Additional arguments ‘…’ are now passed to density.ppp.

• [<-.im
  New argument ‘drop’ controls behaviour when indices are missing
  as in ‘x[] <- value’

• mppm
  Now supports ‘self-starting’ interactions.

• as.im
  New argument ‘stringsAsFactors’ is recognised by many methods.
  It enables the creation of character-string-valued images.

• plot.im
  Axes are now prevented from extending outside the image rectangle.

• plot.im
  New argument ‘zap’.

• blur
  New argument ‘kernel’.

• Smooth.im
  New argument ‘kernel’.

• quadrat.test
  New argument ‘df.est’.

• edge.Ripley
  Numerical stability has been improved on some platforms.
  Results may have changed in marginal cases (e.g. where the circle
  of radius r centred at X is tangent to an edge of the polygonal window).

• rownames, rownames<-, colnames, colnames<-
  These operations now work for hyperframes.

• quadrat.test
  Improved labelling of test when argument ‘CR’ is given.

• plot.pppmatching
  This existing function now has a help file.
  New argument ‘adjust’.

• solist, is.sob, lintess
  Objects of class ‘lintess’ are now recognised as 2D spatial objects
  for the purposes of solist() and is.sob().

• as.linfun.lintess
  Functions created by as.linfun.lintess() now generate better output
  when the function is printed.

BUG FIXES

1. densityVoronoi.lpp
   Did not correctly handle patterns containing duplicated points.
   Fixed.

2. quadrat.test
   Argument ‘CR’ was ignored when method=”MonteCarlo”.
   Fixed.

3. localKcross.inhom, localLcross.inhom
   Argument lambdaX was ignored.
   Fixed.

4. ”[.linim”
   Factor values were erroneously converted to integers, in some cases.
   Fixed.

5. ”[<-.hyperframe”
   Did not handle some classes of objects correctly.
   Fixed.

6. effectfun
   In a multitype point process model, effectfun() ignored any
   user-specified value of ‘marks’.
   Fixed.

7. as.linim.linfun
   Additional arguments (other than spatial coordinates) were ignored.
   Fixed.

8. plot.solist
   Display output was mangled if one entry in the list was
   an object that would normally generate multiple panels of graphics,
   such as an ‘lpp’ or ‘tess’ object with multiple columns of marks,
   or a marked ‘msr’, ‘leverage.ppm’ or ‘influence.ppm’ object.
   Fixed.

9. plot.lpp
   Return value was mangled when x had multiple columns of marks.
   Fixed.

10. colourtable
    Crashed in some cases when ‘breaks’ was given.
    Fixed.

11. rLGCP
    Crashed if ‘win’ was not a rectangle and ‘mu’ was not a constant.
    Fixed.

12. intersect.tess
    Crashed if ‘Y’ was a window object and ‘keepmarks=TRUE’.
    Fixed.

13. envelope.lppm
    Crashed if argument ‘simulate’ was given.
    Fixed.

14. unstack.solist
    Did not correctly handle objects of class ‘lintess’.

15. unstack.solist
    Did not correctly handle objects of class ‘tess’ if they had multiple
    columns of marks.

16. plot.pppmatching
    Issued spurious warnings about unrecognised plot arguments.
    Fixed.

17. plot.lintess
    Issued spurious warnings about unrecognised plot arguments.
    Fixed.

18. shift.lpp, rotate.lpp
    Issued spurious warnings if argument ‘origin’ was given.
    Fixed.

Release notes are available in raw text format here.