1

I'm trying to model a Multidimensional Space-Time Point Pattern (ppx object from spatstat package). But I didn't find any way at package documentation.

I tried this:

library(spatstat)
set.seed(2022)
df <- data.frame(x=runif(4),y=runif(4),t=runif(4),
                 age=rep(c("old", "new"), 2),
                 size=runif(4))
X <- ppx(data=df, coord.type=c("s","s","t","m","m"))
ppm(X~1)

And R showed:

Error in ppm.default(Q = X, trend = ~x, data = NULL, interaction = NULL) : 
  Argument Q must be a point pattern or a quadrature scheme

Is there any way to convert ppx object to point pattern or a quadrature scheme? Should I need do another thing in order to fit a model?

oespinozah
  • 11
  • 2

1 Answers1

1

No. The model-fitting function ppm only accepts point patterns in two-dimensional space (class ppp). Model-fitting for higher dimensional point patterns (class ppx) is not yet supported in the spatstat package.

Adrian Baddeley
  • 2,534
  • 1
  • 5
  • 8