Complex additive models in GEMACO

[Morten Frederiksen]

I have a data set with 4 states and 2 groups. The states consist of 2 physical sites and 2 heterogeneity classes (independently at each site); states 1 and 2 refer to site A, and states 3 and 4 to site B. I now want to construct a model for encounter probability that is additive for heterogeneity state, group and time - but independent for the two sites. In other words, it's a f+g+t model independently for each site. Defining this model in GEMACO has defeated me so far ...

This doesn't work (which doesn't surprise me): firste+nexte.[[f(1,2)+g+t]+[f(3,4)+g+t]]

I've tried a few other rather strange sentences, but nothing has worked so far. Any suggestions?

Morten

[CHOQUET]

Hello morten,

Instead of our sentence which couldn't work, use the following sentence

firste+nexte.[f(1:2).[f+g+t]+f(3:4).[f+g+t]]

A short explanation to that sentence is that you need to apply f+g+t first to f(1:2) and then to f(3:4).
This is achieved by f(1:2).[f+g+t] and f(3:4).[f+g+t]].
For more informations you can read the following article:

Choquet R. (2008). Automatic generation of multistate capture recapture models. The Canadian Journal of Statistics 36(1): 43-57.

Rémi

[Morten Frederiksen]

Thanks Rémi, that does indeed work.

Morten