3

Every time I run model<– coxph(enter,exit, event)~X1+X2+GDP the session crashes suddenly with no error message. It is only when I add the gap variable that it crashes.

I wonder if it has to do with the GDP variable? I tried different sources and it session crashed again. It seems that variables that have higher values such as "223945604003" or "39455684e-5" is the problem?

I use the latest version of R-Studio MAX OS 10.6 (64 bit) on my Mac Yosemite, version 10.10.3

Thank you for all kind of comments.

manlio
  • 18,345
  • 14
  • 76
  • 126
FKG
  • 285
  • 1
  • 4
  • 17

1 Answers1

0

I had the same issue. I didn't have time to debug the coxph source code to see what's going on, but a workaround that worked for me is randomly downsampling the data. I was fitting a model with 1 million observations and getting the session abort, but by reducing it to 100,000 I was able to get it to fit.

Obviously this is highly suboptimal, but it might help somebody get their models fitting, at least.

josliber
  • 43,891
  • 12
  • 98
  • 133