0

I have created an error free ppm in spatstat (R library) together with the associated residuals. It is based around approximately 25,000 data points. However, when I attempt to plot the residuals using plot.msr I get the following message:

Error: all(sigma > 0) is not TRUE

Any suggestions as to what this means and for what I can do to sort out the problem

Thanks

Carl Witthoft
  • 20,573
  • 9
  • 43
  • 73
SimonH
  • 1
  • 1
  • 1
    Hello. Please read the SO recommendations: post a reproducible example of your data and post the code you used to generate residuals, etc. – Carl Witthoft Sep 16 '14 at 14:56
  • Yes, we need more info! It could be that `plot.msr` calls `augment.msr` which in turn calls `Smooth` which may be a good candidate for generating an error like the one you report, but it is impossible to know (and correct) without a reproducible example. – Ege Rubak Sep 16 '14 at 22:11

1 Answers1

0

The original poster (OP) hasn't been very responsive so far, so it is not completely clear what causes this error. For other people encountering this problem a possible workaround may be to call augment.msr on the msr-object with a specific smoothing bandwidth before calling plot.msr. E.g. if x contains the msr-object (such as the residuals of a ppm-object):

x <- augment.msr(x, sigma=s)
plot(x)

where s is a suitable smoothing bandwidth.

Hopefully the problem is fixed in the next version of spatstat (will probably be released in October 2014), but it is impossible to know without assistance from the OP.

Ege Rubak
  • 4,347
  • 1
  • 10
  • 18
  • Apologies for the delay. Using the 'augment.msr' command together with the latest version of spatstat does indeed solve the problem. Thanks – SimonH Jan 14 '15 at 12:09