09 January 2019
We thank Andrew Bevan, Hamidreza Heydarian and Andrew P Webster for contributions.
Mark-mark scatter plot.
Standard error calculation for inverse-distance weighting
Minor improvements and extensions.
Version nickname: ‘Drop Bear’
SIGNIFICANT USER-VISIBLE CHANGES
Standard errors can now be calculated by setting ‘se=TRUE’.
Computation accelerated, especially when NA’s are absent.
New arguments ‘fun.handles.na’ and ‘check’.
kppm, dppm, clusterfit
Changed precedence rule for handling the algorithm parameters
in the minimum contrast algorithm. Individually-named arguments
‘q,p,rmax,rmin’ now take precedence over entries
with the same names in the list ‘ctrl’.
Improved algorithm stability
Prints more information about algorithm convergence.
Argument ‘what’ can take the value “ijd”
Graphical parameters, specified when the object was created,
are now taken as the defaults for graphical parameters to the plot.
When result=”im” the resulting image did not have the same dimensions
as the input. [Spotted by Hamidreza Heydarian.]
If argument ‘origin’ was given, and if result=”im” was specified,
the resulting image was wrongly displaced.
Result was mangled when n=0 or n=1.
Release notes are available in raw text format here.