Is it possible to plot density surfaces using the groups argument?

[Milena Ribeiro]

Hello, 

When using the "groups" argument in secr.fit, is it possible to plot distinct density surfaces for each of the groups, i.e. one density surface for each sex (F and M)?

I am using the following command:
surface.sex <- predictDsurface(fit.sex.D)
plot(surface.sex)

And I get the error:
Error in plot.Dsurface(surface.sex) : covariate D.0 not found

I understand that when I use the "groups" argument in secr.fit(), secr fits separate models for each group and stores the density estimates separately. However, I do not know how to solve the problem in the script and if it is possible to solve it.

Thank you for your attention,

Milena Ribeiro.

[Murray Efford]

For groups, density predictDsurface() saves density as one covariate for each group. If the groups are 'f' and 'm' then these are named 'D.f' and 'D.m'. 

To plot a particular group 'f' you can then

(i) specify e.g. plot(surface.sex, 'D.f') where the second argument is the covariate name, or more directly 

(ii) specify e.g. plot(surface.sex, group = 'f').  The group argument is documented in  ?plot.Dsurface.

Murray

# I tested with:

library(secr)
morning <- subset(housemouse, occ = c(1,3,5,7,9)) # includes one mouse with sex unknown
groupx   <- secr.fit(morningx, model = D ~ g + x, groups = 'sex', buffer = 25, biasLimit = NA, trace = FALSE)
Ds <- predictDsurface(groupx)
par(mfrow=c(1,2))

plot(Ds, group = 'f')
plot(Ds, group = 'm')

[Milena Ribeiro]

Hello, Efford,

Thank you for your response. I tried what you recommended, as well as some variations for testing purposes, and in all cases I received the following error:

r

CopyEdit

Error in seq.default(covrange[1], covrange[2], length.out = breaks + 1): 'from' must be a finite number In addition: Warning messages: 1: In min(x): no non-missing arguments to min; returning Inf 2: In max(x): no non-missing arguments to max; returning -Inf

When I run the command covariates(surface.sex), I get the columns corresponding to my groups, but all values in the rows (i.e., the mask points) are NA, which I believe is what’s causing the error above. Is there any solution to this?

Thank you again for your help.

Best regards,
Milena Ribeiro

[Murray Efford]

So predictDsurface is not working for your model and mask, for some reason we cannot see. I think you will need to troubleshoot that, looking at the model and the fitted coefficients coef( fit.sex.D) and the covariates of the original mask covariates(fit.sex.D$mask). It worked for me, as you can see by running the example.

Murray