[R] nlme: fixed=lKe+lKa+lV~1 fails, fixed=lV+lKe+lKa~1, works

[Dieter Menne]

Dear Douglas Bates,

Looks like the success of nlme depends on the order of the fixed coefficients.
In a simple variation of your Quinidine-model (with simulated data,
result is known), using the following call give a reasonable result:

Orli12.nlme<-nlme(
     conc~orliModel(Pat,time,conc,dose,interval,lV,lKa,lKe),
     data=Orli12,
     fixed=lV+lKe+lKa~1,
     random=lV~1,
     groups=~Pat,
     start=list(fixed=c(log(V),log(Ka),log(Ke))),
     na.action=NULL,
     naPattern=~!is.na(conc)
     )

Nonlinear mixed-effects model fit by maximum likelihood
  Model: conc ~ orliModel(Pat, time, conc, dose, interval, lV, lKa, lKe)
  Data: Orli12
  Log-likelihood: -1016
  Fixed: lV + lKe + lKa ~ 1
    lV    lKe    lKa
-5.796 -1.628 -0.483

Random effects:
 Formula: lV ~ 1 | Pat
           lV Residual
StdDev: 0.097     20.3


With
     fixed=lKe+lKa+lV~1,

I get reproducably:

Error in nlme.formula(conc ~ orliModel(Pat, time, conc, dose, interval,  :
        Step halving factor reduced below minimum in PNLS step


---------------------------------------
Dr. Dieter Menne
Biomed Software
72074 Tübingen
Tel (49) (7071) 52176
FAX (49) (7071) 55 10 46
dieter.menne at menne-biomed.de
www.menne-biomed.de

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._