time-varying individual covariates with known fate models
Posted: Wed Dec 06, 2017 3:12 pm
Hi, I try to test the effect of the normalized difference vegetation index (ndvi) on survival of migratory caribou for different seasons of activity. All caribou are equiped with radio-telemetry. I am working with 12 years of data and I built an inp file for each different season. Each year I have 3 different numeric time-varying individual covariates.
That's a small part of the data after importation in Rmark.
I follow the indication I found on the appendixC - chapter 16 on individuals covariates and I named each covariate as a concatenation of the variable and the time occasion.
For example I have 12 years so time occasion going to 1 to 12 and the first covariate is named "jn1", "jn2", ..., "jn12". So I expected to be able to build models like :
But Rmark reject it and send me these error-messages :
What I understand is that I've been unable to insert these covariates in the design data even if when I imported the data I specify that those were covariates. I also see them in the output of the structure of the process.data.
But I don't understand where I made a mistake. Is it possible to test the effect of a non factoriel time-variable covariates with known-fate models ? I couldn't find similar example of what I'm trying to do.
Could you help me ? If you need more information about my script or the data base, just let me know. Thank you.
Barbara
That's a small part of the data after importation in Rmark.
- Code: Select all
ch freq jn1 jl1 p1 jn2 jl2 p2 jn3 jl3 p3
2004001 000000000010101000000000 1 0.058 0.23 0.255 0.146 0.335 0.114 -0.025 0.233 0.292
2004002 000000000010101010100000 1 0.058 0.23 0.255 0.146 0.335 0.114 -0.025 0.233 0.292
2004004 000000000010101000000000 1 0.058 0.23 0.255 0.146 0.335 0.114 -0.025 0.233 0.292
2005002 000000000010101010101100 1 0.058 0.23 0.255 0.146 0.335 0.114 -0.025 0.233 0.292
2005004 000000000010101000000000 1 0.058 0.23 0.255 0.146 0.335 0.114 -0.025 0.233 0.292
2005005 000000000010001010100000 1 0.058 0.23 0.255 0.146 0.335 0.114 -0.025 0.233 0.292
...
I follow the indication I found on the appendixC - chapter 16 on individuals covariates and I named each covariate as a concatenation of the variable and the time occasion.
For example I have 12 years so time occasion going to 1 to 12 and the first covariate is named "jn1", "jn2", ..., "jn12". So I expected to be able to build models like :
- Code: Select all
S.time_JN=list(formula=~time:jn)
But Rmark reject it and send me these error-messages :
Error in make.mark.model(data.proc, title = title, parameters = model.parameters, :
Error: Variable jn used in formula is not defined in data
Error in mark(model.parameters = model.parameters, initial = initial, :
Misspecification of model or internal error in code
What I understand is that I've been unable to insert these covariates in the design data even if when I imported the data I specify that those were covariates. I also see them in the output of the structure of the process.data.
But I don't understand where I made a mistake. Is it possible to test the effect of a non factoriel time-variable covariates with known-fate models ? I couldn't find similar example of what I'm trying to do.
Could you help me ? If you need more information about my script or the data base, just let me know. Thank you.
Barbara