[FLR-list] FLSR sigma2

[Laurie]

Probably why for the last two years a Britain has won the Tour De France!
Laurie

On 08/08/13 13:59, Clara Ulrich wrote:
> Ok thanks for quick answer!
>
> And I don't know why but I still find it easier to remember how to bike - probably because bike's design and functioning hasn't changed that much from when I was a child, much less new fancy features popping up;-)
> Clara
>
> -----Original Message-----
> From: flr-list-bounces at r-forge.wu-wien.ac.at [mailto:flr-list-bounces at r-forge.wu-wien.ac.at] On Behalf Of Iago MOSQUEIRA (JRC)
> Sent: 08 August 2013 13:36
> To: flr-list at r-forge.wu-wien.ac.at
> Subject: Re: [FLR-list] FLSR sigma2
>
> On 08/08/2013 12:19 PM, Clara Ulrich wrote:
>> Hi all,
>>
>> I feel that my FLR brain is smelting away, I just do not use it enough anymore to keep a sharp up-to-date knowledge ;-( And indeed, trying to re-run some 2 years old from R2.8.1 I get some bugs of course.
>>
>> Is there nothing like a sigma2 in the FLSR parameters anymore? Is this stored something else?
>>
>> Thanks
>> Clara
> Hi,
>
> But FLRing is like cycling, never to be forgotten...
>
> The formulation of the likelihood that has been used for the last 2-3 years does not have sigma2 as an estimated parameter, but is instead calculated, as it was just a nuisance parameter that made life more dificult forn the optimizer
>
>> getMethod('loglAR1', c('FLQuant', 'FLQuant'))
>
> Method Definition:
>
> function (obs, hat, ...)
> {
>      .local <- function (obs, hat, rho = 0)
>      {
>          n <- dim(obs)[2]
>          rsdl <- (obs[, -1] - rho * obs[, -n] - hat[, -1] + rho *
>              hat[, -n])
>          s2 <- sum(rsdl^2, na.rm = T)
>          s1 <- s2
>          if (!all(is.na(rsdl[, 1])))
>              s1 <- s1 + (1 - rho^2) * (obs[, 1] - hat[, 1])^2
>          sigma2 <- sigma(obs, hat)^2
>          n <- length(obs[!is.na(obs)])
>          sigma2.a <- (1 - rho^2) * sigma2
>          res <- (log(1/(2 * pi)) - n * log(sigma2.a) + log(1 -
>              rho^2) - s1/(2 * sigma2.a))/2
>          if (!is.finite(res))
>              res <- -1e+100
>          return(res)
>      }
>      .local(obs, hat, ...)
> }
> <environment: namespace:FLCore>
>
> which class sigma()
>
>> getMethod('sigma', c('FLArray'))
> Method Definition:
>
> function (object, ...)
> {
>      .local <- function (object, hat = rep(0, length(object)))
>      {
>          if (all(is.na(hat)))
>              return(Inf)
>          SS <- sum((object - hat)^2, na.rm = TRUE)
>          return((SS/length(hat[!is.na(hat)]))^0.5)
>      }
>      .local(object, ...)
> }
> <environment: namespace:FLCore>
>
> Signatures:
>          object
> target  "FLArray"
> defined "FLArray"
>
> which calculates it from SS.
>
>
> Iago
>
>
>
> --
> Dr Iago Mosqueira
>
> European Commission - Joint Research Center Institute for the Protection and Security of the Citizen (IPSC) Maritime Affairs Unit FISHREG - Scientific Support to Fisheries TP 051, Via Enrico Fermi 2749
> I-21027 Ispra (VA), Italy
>
> Office : +39 0332 785413
> Fax: +39 0332 789658
>
> iago.mosqueira at jrc.ec.europa.eu
>
> http://fishreg.jrc.ec.europa.eu/home
>
> _______________________________________________
> flr-list mailing list
> flr-list at flr-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/flr-list
> _______________________________________________
> flr-list mailing list
> flr-list at flr-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/flr-list