[R] Parameters setting in functions optimization
Diane Bailleul
diane.bailleul at u-psud.fr
Wed Nov 30 17:18:00 CET 2011
Dear Florent,
I know that I'm asking to optim to minimize my values, and that the
results with a lower fvalue are best supported than those with a higher
fvalue.
My comment was just from a data point of view. I'd like the lower ms
(second parameter) as possible, as well as the fvalue. So a ms of 0.97
(i.e. that 97% of my individuals are coming from outside the experiment
plot) is very disapointing.
Dear John,
I can send you my data by email. It's very kind of you to offer to use
my data as test for your new and coming optimx ! thank you.
Le 11/30/2011 12:08 PM, Florent D. a écrit :
> With optimx(c(30,50),ms=c(0.4,0.5), fn=LogLiketot)
>
> where
>
> LogLiketot<- function(dist,ms) {
> res<- NULL
> for(i in 1:nrow(pop5)) {
> for(l in 1:nrow(freqvar)) {
> res<- c(res, pop5[i,l]*log(LikeGi(l,i,dist,ms)))
> }
> }
> return(-sum(res))
> }
>
> I think it will do something like this for the first call to LogLiketot:
>
> LogLiketot(c(30,50), ms=c(0.4,0.5))
>
> which is obviously not the usage you had in mind.
>
> Also, I see you say the results for the bad usage above:
>
> par fvalues method fns grs itns conv KKT1 KKT2 xtimes
>> 2 19.27583, 25.37964 2249.698 BFGS 12 8 NULL 0 TRUE TRUE 57.5
>> 1 29.6787861, 0.1580298 2248.972 Nelder-Mead 51 NA NULL 0 TRUE TRUE 66.3
> look very good but you do not comment about the results for the
> correct usage of optimx:
>
> par fvalues method fns grs itns conv KKT1 KKT2 xtimes
>> 2 39.9969607, 0.9777634 1064.083 BFGS 29 10 NULL 0 TRUE NA 92.03
>> 1 39.7372199, 0.9778101 1064.083 Nelder-Mead 53 NA NULL 0 TRUE NA 70.83
> Do you realize optimx is trying to _minimize_ your function? See that
> the fvalues from the correct usage are much better (smaller) than you
> first (bad) usage.
>
>
>
>
>
> On Wed, Nov 30, 2011 at 4:16 AM, Diane Bailleul
> <diane.bailleul at u-psud.fr> wrote:
>> Le 11/30/2011 2:09 AM, Florent D. a écrit :
>>
>> Thanks for your answer !
>>
>>> I also think your last write-up for LogLiketot (using a single
>>> argument "par") is the correct approach if you want to feed it to
>>> optim().
>> I'm not dedicated to optim() fonction. I just want to optimise my two
>> parameters and the loglikelihood result, and if there's a better fonction
>> for that, I wish I could use it.
>>
>>
>>> So now you have a problem with log(LikeGi(l, i, par[1], par[2])) for
>>> some values of par[1] and par[2].
>>>
>>> Where is LikeGi coming from? a package or is it your own function?
>> My own function, otherwise it would be simplier to discuss about my
>> problems.
>>
>>
>>> You could add some print statements (if you are familiar with
>>> "browser()" it is even better) so you may see what values of "par" are
>>> causing trouble.
>> I'm not familiar, but I'll search about browser().
>>
>> If the function with par is correct, any idea of what I've made with this :
>>
>> optimx(c(30,50),ms=c(0.4,0.5), fn=LogLiketot)
>>
>>
>> ?
>>
>>
>>
>>>
>>> On Tue, Nov 29, 2011 at 1:15 PM, Diane Bailleul
>>> <diane.bailleul at u-psud.fr> wrote:
>>>> Good afternoon everybody,
>>>> I'm quite new in functions optimization on R and, whereas I've read lot's
>>>> of
>>>> function descriptions, I'm not sure of the correct settings for function
>>>> like "optimx" and "nlminb".
>>>> I'd like to minimize my parameters and the loglikelihood result of the
>>>> function.
>>>> My parameters are a mean distance of dispersion and a proportion of
>>>> individuals not assigned, coming from very far away.
>>>> The function LikeGi reads external tables and it's working as I want
>>>> (I've
>>>> got a similar model on Mathematica).
>>>>
>>>> My "final" function is LogLiketot :
>>>> LogLiketot<- function(dist,ms)
>>>> {
>>>> res<- NULL
>>>> for(i in 1:nrow(pop5)){
>>>> for(l in 1:nrow(freqvar)){
>>>> res<- c(res, pop5[i,l]*log(LikeGi(l,i,dist,ms)))
>>>> }
>>>> }
>>>> return(-sum(res))
>>>> }
>>>>
>>>> dist is the mean dispersal distance (0, lots of meters) and ms the
>>>> proportion of individuals (0-1).
>>>> Of course, I want them to be as low as possible.
>>>>
>>>> I'd tried to enter the initials parameters as indicated in the tutorials
>>>> :
>>>> optim(c(40,0.5), fn=LogLiketot)
>>>>> Error in 1 - ms : 'ms' is missing
>>>> But ms is 0.5 ...
>>>>
>>>> So I've tried this form :
>>>> optimx(c(30,50),ms=c(0.4,0.5), fn=LogLiketot)
>>>> with different values for the two parameters :
>>>> par fvalues method fns grs itns conv KKT1 KKT2
>>>> xtimes
>>>>> 2 19.27583, 25.37964 2249.698 BFGS 12 8 NULL 0 TRUE TRUE
>>>>> 57.5
>>>>> 1 29.6787861, 0.1580298 2248.972 Nelder-Mead 51 NA NULL 0 TRUE TRUE
>>>>> 66.3
>>>> The first line is not possible but as I've not constrained the
>>>> optimization
>>>> ... but the second line would be a very good result !
>>>>
>>>> Then, searching for another similar cases, I've tried to change my
>>>> function
>>>> form:
>>>>
>>>> LogLiketot<- function(par)
>>>> {
>>>> res<- NULL
>>>> for(i in 1:nrow(pop5)){
>>>> for(l in 1:nrow(freqvar)){
>>>> res<- c(res, pop5[i,l]*log(LikeGi(l,i,par[1],par[2])))
>>>> }
>>>> }
>>>> return(-sum(res))
>>>> }
>>>>
>>>> where dist=par[1] and ms=par[2]
>>>>
>>>> And I've got :
>>>> optimx(c(40,0.5), fn=LogLiketot)
>>>> par fvalues method fns grs itns conv KKT1 KKT2
>>>> xtimes
>>>>> 2 39.9969607, 0.9777634 1064.083 BFGS 29 10 NULL 0 TRUE NA
>>>>> 92.03
>>>>> 1 39.7372199, 0.9778101 1064.083 Nelder-Mead 53 NA NULL 0 TRUE NA
>>>>> 70.83
>>>> And I've got now a warning message :
>>>>> In log(LikeGi(l, i, par[1], par[2])) : NaNs produced
>>>> (which are very bad results in that case)
>>>>
>>>>
>>>> Anyone with previous experiences in optimization of several parameters
>>>> could
>>>> indicate me the right way to enter the initial parameters in this kind of
>>>> functions ?
>>>>
>>>> Thanks a lot for helping me !
>>>>
>>>> Diane
>>>>
>>>> --
>>>> Diane Bailleul
>>>> Doctorante
>>>> Université Paris-Sud 11 - Faculté des Sciences d'Orsay
>>>> Unité Ecologie, Systématique et Evolution
>>>> Département Biodiversité, Systématique et Evolution
>>>> UMR 8079 - UPS CNRS AgroParisTech
>>>> Porte 320, premier étage, Bâtiment 360
>>>> 91405 ORSAY CEDEX FRANCE
>>>> (0033) 01.69.15.56.64
>>>>
>>>> ______________________________________________
>>>> R-help at r-project.org mailing list
>>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>>> PLEASE do read the posting guide
>>>> http://www.R-project.org/posting-guide.html
>>>> and provide commented, minimal, self-contained, reproducible code.
>>>>
>>
>> --
>> Diane Bailleul
>> Doctorante
>> Université Paris-Sud 11 - Faculté des Sciences d'Orsay
>> Unité Ecologie, Systématique et Evolution
>> Département Biodiversité, Systématique et Evolution
>> UMR 8079 - UPS CNRS AgroParisTech
>> Porte 320, premier étage, Bâtiment 360
>> 91405 ORSAY CEDEX FRANCE
>> (0033) 01.69.15.56.64
>>
>>
--
Diane Bailleul
Doctorante
Université Paris-Sud 11 - Faculté des Sciences d'Orsay
Unité Ecologie, Systématique et Evolution
Département Biodiversité, Systématique et Evolution
UMR 8079 - UPS CNRS AgroParisTech
Porte 320, premier étage, Bâtiment 360
91405 ORSAY CEDEX FRANCE
(0033) 01.69.15.56.64
More information about the R-help
mailing list