On 07/06/17 22:22, Seth Flaxman wrote:

> I've got the intensity of an inhomogeneous Poisson process (fit using

> some new methods I'm working on, so not created by spatstat) as an

> image object, the observed point pattern as a ppp object, and I'd like

> to call residuals.ppm to compute residuals. Is there a simple way to

> create a fitted point process model (ppm) from an image and points so

> I can pass this directly to residuals.ppm?

Sorry for the long delay in responding.

There are two ways to do this:

[1] Fit a model using the putative intensity as an offset:

fit1 <- ppm(X ~ offset(log(lam)))

where 'lam' is the pixel image of intensity, and 'X' is the point pattern.

Then do:

res <- residuals(fit1)

[2] Fit some other model (e.g. just a constant intensity model) and use

the argument 'fittedvalues' of residuals.ppm to specify the fitted

intensity values.

fit2 <- ppm(X ~ 1, forcefit=TRUE)

res <- residuals(fit2, fittedvalues=lam[quad.ppm(fit2)])

cheers,

Rolf Turner

--

Technical Editor ANZJS

Department of Statistics

University of Auckland

Phone: +64-9-373-7599 ext. 88276

_______________________________________________

R-sig-Geo mailing list

[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo