Utilization Distributions

[Mackenzie Jones]

Any idea why this looks different?

[Mackenzie Jones]

For more information, this is the error message that is occurring:

Warning messages:
1: In cov.loglike(hess, grad) :
  MLE is near a boundary or optimizer failed.
2: In ctmm.fit(ds.tel.30, GUESS, method = "pHREML") :
  pREML failure: indefinite ML Hessian or divergent REML gradient.
3: In cov.loglike(DIFF$hessian, grad) :
  MLE is near a boundary or optimizer failed.

[Christen Fleming]

Hi Mackenzie,

Those are warning messages (rather than error messages). Are you running ctmm.fit rather than ctmm.select ?

Then if you do run ctmm.select, is the selected model IID?

And if the selected model is IID, are there large errors or tiny sampling intervals in the data that need to be dealt with?

Best,

Chris