1 reply to this topic

[raghava choudary]

Dear Si mon,

 

I am facing a problem with FO estimates in phoenix NLME.

 

Previously i forwarded a dataset to you and the values are as those in the attached file (IV_PO simon 29Jan2011.phxproj)

 

Now i analyzed the same dataset and i am getting different FO estimates, for V2 and CL2.

(IV-PO230511 rat.phxproj)

 

Looking forward for your reply,

 

Thanks with best regards,

 

raghav [file name=IV_PO_230511_rat.phxproj size=650688]http://pharsight.com/extranet/media/kunena/attachments/legacy/files/IV_PO_230511_rat.phxproj[/file]

[Simon Davis]

HI Raghav, I am travelling and training this week so I only took the brief-est look at this, - if you look at you more recent model code you have left all initial estimates at 1, if you run it with the same initial estimates as previously you will get the same results I think;

 

fixef(tvF = c(, 0.4, ))

    fixef(tvV = c(, 1.5, ))

    fixef(tvCl = c(, 1.4, ))

    fixef(tvKa = c(, 1.4, ))

    fixef(tvV2 = c(, 1, ))

    fixef(tvCl2 = c(, 4.5, ))

 

I guess that numerically you're falling into some local minimum and not finding an optimum solution - I think one of the other algorithms may help you with that. In future it would help to be more specific with your problem so we know the differences you're talking about before exploring around the project.

 

 Simon.