FLR

FS#32: as.data.frame(FLQuants(m(ple4))) name of data is "missing"

Laurie Kell — 2010-08-09T13:44:41Z

head(as.data.frame(FLQuants(m(ple4))))

name of data is "missing"

FS#31: fixed in FLGrowth

Laurie Kell — 2010-07-31T11:08:39Z

Looks like you cant fix parameters in FLGrowth

# class FLGrowth
setClass('FLGrowth', representation('FLModel',
mass='FLArray'))

# constructor
setGeneric('FLGrowth', function(model, ...)
standardGeneric('FLGrowth'))

setMethod('FLGrowth', signature(model='ANY'),
function(model, ...)
return(FLModel(model, ..., class='FLGrowth')))

setMethod('FLGrowth', signature(model='missing'),
function(...)
return(FLModel(formula(NULL), ..., class='FLGrowth')))

# example
data(ple4)

gro <- FLGrowth(mass~(Minf*(1-exp(-k*(age-t0))))^3, mass=FLCohort(stock.wt(ple4)))

gro <- fmle(gro, start=list(Minf=7, k=0.3, t0=0), fixed=list(Minf=7))

FS#30: FLCohort constructor not fully functional

Laurie Kell — 2010-07-30T08:33:01Z

This works

FLQuant(0:20,dimnames=list(age=0:20,year=1980:2010))

so why not?

FLCohort(0:20,dimnames=list(age=0:20,cohort=1980:2010))

various other things missing from FLCohort

FS#29: ple4["10"]

Laurie Kell — 2010-07-27T09:31:57Z

this works

ple4[10]

but this fails

ple4[ac(10)]

FS#28: FLPar bugs

Laurie Kell — 2010-07-26T15:33:38Z

#I´m using Version: 2.3-644 and I´m having problems with FLPar. #These dont appear to work but will be required for many applications FLPar(c(Linf=238.6,K=0.185,t0=-1.404),params=c("Linf","K","t0"),unit=c("male","female"),iter=1) FLPar(c(Linf=238.6,K=0.185,t0=-1.404),params=c("Linf","K","t0"),season=1:4,iter=1) FLPar(c(Linf=238.6,K=0.185,t0=-1.404),params=c("Linf","K","t0"),areas=1:2,iter=1) #This works params <-FLPar(array(c(Linf=238.6,K=0.185,t0=-1.404),c(3,2,4,1),list(params=c("Linf","K","t0"),unit=c("male","female"),season=1:4,iter=1))) ## but show is wrong params params@.Data ## this doesn´t work params["Linf","male",,]<-params["Linf","female",,]*.75 ## this does params@.Data["Linf","male",,]<-params@.Data["Linf","female",,]*.75

FS#27: as.data.frame(FLQuant) doesn´t return the same as as.data.frame(FLQuant)

Laurie Kell — 2010-07-26T10:14:42Z

as.data.frame(FLQuant) doesn´t return the same as as.data.frame(FLQuant) as in FLQuants the data.frame returns factors, can we make these the same. mapply(function(x,y) class(y[x]), x=names(t.),y=as.data.frame(m(ple4))) age year unit season area iter data qname "numeric" "numeric" "factor" "factor" "factor" "factor" "numeric" "numeric" mapply(function(x,y) class(y[x]), x=names(t.),y=as.data.frame(FLQuants(m(ple4)))) age year unit season area iter data qname "factor" "numeric" "factor" "factor" "factor" "factor" "numeric" "character" If they have to be factors can we also make sure that for ages as years teh order matches the numeric order Laurie

FS#26: using mcf loses names

Laurie Kell — 2010-03-13T10:02:41Z

using mcf loses names

head(as.data.frame(FLQuants(m=m(ple4))))
age year unit season area iter data qname
m.1 1 1957 unique all unique 1 0.1 m
m.2 2 1957 unique all unique 1 0.1 m
m.3 3 1957 unique all unique 1 0.1 m
m.4 4 1957 unique all unique 1 0.1 m
m.5 5 1957 unique all unique 1 0.1 m
m.6 6 1957 unique all unique 1 0.1 m
> head(as.data.frame(mcf(FLQuants(m=m(ple4)))))
age year unit season area iter data qname
NA.1 1 1957 unique all unique 1 0.1
NA.2 2 1957 unique all unique 1 0.1
NA.3 3 1957 unique all unique 1 0.1
NA.4 4 1957 unique all unique 1 0.1
NA.5 5 1957 unique all unique 1 0.1
NA.6 6 1957 unique all unique 1 0.1

FS#25: FLCohort; bugs and missing bits

Laurie Kell — 2010-02-17T09:15:57Z

There are a few bugs in FLCohort that make it difficult to use i.e.

#1) There is no creater
flc <-FLCohort(NA,dimnames=list(age=1:10,year=1980:2000,iter=1:100))

# You can get round this by creating an FLQuant first but that is not what you need if you want to create year-class structure
# e.g. fast/slow growing cohorts
flc <-FLCohort(FLQuant(NA,dimnames=list(age=1:10,year=1980:2000,iter=1:100)))

#2) You can´t create an FLQuant from an FLCohort
flq <-FLQuant(flc)

#You can work around this by creating a data.frame first
df <-as.data.frame(flc)
df$year<-df$cohort+df$age

#3) but even though you have all the required columns in te data.frame you can´t create the FLQuant
flq <-as.FLQuant(df)

# You have to remove the redundant column, I don´t see why this be so. The redundant data should be ignored
flq <-as.FLQuant(df[,-2])

#4) There is no corresponding as.FLCohort
flc <-as.FLCohort(df[,-8])

#5) You can´t create a data.frame from an FLCohort
dfc<-data.frame(flc)

We can correct this by implementing the corresponding methods

Laurie

FS#24: Changes to FLSR to make it more robust and consistent

Laurie Kell — 2009-12-23T09:28:27Z

Changes to FLSR to make it more robust and consistent

1) Change algorithm for setting initial values in BH
2) Add limits to segreg to ensure breakpoint doesn´t fall outside of the data
3) Calculate sigma from the residuals rather than fitting it as part of the likelihood
4) Add auto parscaling
5) Change all likelihoods to lognormal or at least put a option to choose
6) Ensure residuals are on the log scale if a log normal error model is used.
7) turn sv2ab and ab2sv into methods
8) check that steepness & virgin biomass options work across all classes

1,2,3 and 4 then Beverton and Holt, Ricker and segmented regression SRRs can be fitted to the ICES stock database. Otherwise fits can fail which can cause problems in that
parameters are wrong especially for Monte Carlo simulations.

Problems are due to FLSR simply not being able to find the solution, there are several reasons for this, poor initial values, going off into parameter space with a flat likelihood, different scales for parameters

In some methods normal log likelihood is used while in other the log normal, e.g. in bevholt.ar1() it is the log normal as this is needed to deal with AR1 processes, but you can fixed rho to 0 and you then have a standard normal likelihood but this may differ from bevholt()

Diagnostic plots are on the untransformed scale even when the error model is lognormal, this is because residuals are not on the log scale, this means that the QQ plots are wrong, the residual plots misleading and difficult to interpret and that passing FLSR objects to other methods (e.g. fwd for projection) liable to error.

Most of these suggestions have been coded up and are in FLBocadillio, main task is to check them and add to FLCore

FS#23: Iters

Laurie Kell — 2009-10-30T17:43:10Z

1) Secondary classes don´t always work correctly when mixing iters as 1
or n, for example having landings.sel with n iters in FLBRP means that
landings.n must also have n iters. I had major problems in fwd() when
working with FLStocks, FLBiols, FLMetiers, FLFleets, FLCatches and FLSRs
and so created some validation code, found in FLCoreVarCon.R. I think it
would be worthwhile either adding this to FLCore or creating some
protocols for use in secondary packages.

I therefore propose we look at FLCoreVarCon.R in FLBRP and add it to FLCore as a generic routine
for use in secondary packages