Alex Ocampo

icon indicating an email [email protected]

icon company Novartis icon location Switzerland icon location Ph.D. Biostatistics, Harvard University, 2020 B.S. Statistics, University of Michigan, 2012

Results 2 comments of Alex Ocampo

numerical overflow in calculating likelihood

I'm also running into the same issue. It is definitely an identifiability issue as Prof Jackson eloquently described above. I believe I can handle this in an ad hoc manner...

numerical overflow in calculating likelihood

That makes sense. Also, (facepalm)... of course they have more parameters. Thanks!