degrees of freedom in LL computation

  • Roberto



    I’m developing a model (non-linear and with stiff ODE’s) and when running LL estimation (Imp. Sampling) with its different (and very similar) variants and with dof optimized between 5, 10 and 15, I see that the number of dof chosen by the algorithm changes in the different variants. If I want to compare the model variants in terms of LL, should I fix the # of dof or can I compare LL values obtained with different # of dof?

    Thank you,
    Roberto Bizzotto

  • Cocas



    I have the same question.
    Also, regarding the standard error of the LL is there any threshold value that we should be looking for?

    Thank you,
    Hélder Duarte

Viewing 2 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic.

Robertodegrees of freedom in LL computation