Share your ideas for the JMP Scripting Unsession at Discovery Summit by September 17th. We hope to see you there!
Choose Language Hide Translation Bar
Highlighted
bayesfactor
Level III

How to resolve convergence issues in Fit Parametric Survival platform?

I'm using the Fit Parametric Survival platform and finding convergence issues with many models that I run. For example, I can use 1 model effect and get convergence, and sometimes I can use 2 model effects (independent variables), but other combinations of 2 model effects will not converge.  I don't think I have too many degrees of freedom -- one model that doesn't converge has 11 DF, while another model that does converge (on the same data) has 49DF.  What tips are there to address convergence issues? 

5 REPLIES 5
Highlighted

Re: How to resolve convergence issues in Fit Parametric Survival platform?

What distribution are you using to model the errors? What does a plot of the residuals look like?

Learn it once, use it forever!
Highlighted
bayesfactor
Level III

Re: How to resolve convergence issues in Fit Parametric Survival platform?

I'm using a Weibull. Here are the residuals:

Untitled.png

Highlighted

Re: How to resolve convergence issues in Fit Parametric Survival platform?

How much of the data is censored? What kind of censoring happens with your data?

Learn it once, use it forever!
Highlighted
bayesfactor
Level III

Re: How to resolve convergence issues in Fit Parametric Survival platform?

61/653 samples are right-censored. 

Highlighted

Re: How to resolve convergence issues in Fit Parametric Survival platform?

Nothing so far seems unusual or likely the cause for non-convergence. I recommend that you contact JMP Technical Support at support@jmp.com and ask for help. We encourage users to seek help here but your problem requires technical support.

Learn it once, use it forever!
Article Labels

    There are no labels assigned to this post.