[R] Help about nlminb function

Ben Bolker bbolker at gmail.com
Wed Dec 15 15:47:26 CET 2010


kamel gaanoun <kamel.gaanoun <at> gmail.com> writes:

> 
> I appologize for the lack of information in my previous mail, but it is my
> first one.
> 
> Yes  I red -in part- the Port documentation to learn what does the message
> X-convergence (3) mean.
> 
> so this is my nlminb and the results :
> 
> res1<-nlminb(parametres[,3], V, lower=c(rep(0.01, 18), rep(0.01, 3),
> rep(-Inf, 6)), upper=c(rep(Inf, 18), rep(0.99, 3), rep(Inf, 6)), control =
> list(maxit=1000) )
> 
> Results : > res1
> $par
>  [1] 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000
>  [8] 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000 1.0000000
> [15] 1.0000000 1.0000000 1.0000000 1.0000000 0.6899225 0.7993528 0.6480687
> [22] 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000

  The very most likely scenario is that you messed up your objective function
somehow. I do this on a regular basis despite many years of experience.
Test your objective function with your starting value (it should give
a finite value), with values near your starting value (it should give
similar answers), and with parameter values that you suspect to be worse
(but not crazily worse) (it should give higher values).

  Ben Bolker



More information about the R-help mailing list