[R-sig-Geo] problem with ppm.plot in spatstat since version 1.15

Jarek Jasiewicz jarekj at amu.edu.pl
Wed Mar 10 22:17:27 CET 2010


Hi!

I encountered following problem with R spatstat with plot.ppm trying 
reporduce some trend analysis on my old data:

to reproduce it just enough:

x1=runif(100)
y1=runif(100)

e=1
w=0
s=0
n=1

x=x1*(e-w)+w
y=y1*(n-s)+s

wind=owin(c(w,e),c(s,n))
pp=ppp(x,y,window=wind)
pp_m=ppm(pp, ~polynom(x,y,2))
plot(pp_m)

everything is OK!

but if coordinates are greater for example:
e=10000
w=1000
s=1000
n=10000

x=x1*(e-w)+w
y=y1*(n-s)+s

wind=owin(c(w,e),c(s,n))
pp=ppp(x,y,window=wind)
pp_m=ppm(pp, ~polynom(x,y,2))
plot(pp_m)


Error occured:

Error in solve.default(fisher) :
  system is computationally singular: reciprocal condition number = 
4.33811e-18
Error in mmi %*% vc : requires numeric/complex matrix/vector arguments


Till version 1.15 of spastat and R 2.8 t works!

Thanks for help
Jarek



More information about the R-sig-Geo mailing list