hzar.doCLTData1DRaw: Create a 'hzar.obsData' object using a table of individual...
In GrahamDB/hzar: Hybrid Zone Analysis using R

Description Usage Arguments Details Value Author(s) See Also Examples

Create a hzar.obsData object using a table of individual traits.

1
2
3

hzar.doCLTData1DRaw(distance, traitValue)
hzar.doNormalData1DRaw(site.dist, traitSite, traitValue)
hzar.mapSiteDist(siteID, distance)

`distance`	The distance of the sampling site. For `hzar.doCLTData1DRaw`, samples at the same distance are treated as being from the same sampling site.
`traitValue`	The value of the trait of the individual sampled.
`traitSite`	The id of site where the individual was found.
`site.dist`	A named vector mapping site id codes to the distance of the sampling site. The function `hzar.mapSiteDist` returns a suitable vector.
`siteID`	The list of id codes associated with the sampling site. This list should be identical in length to `distance`, each entry must be unique, and the order of the sites referenced must be identical for `distance` and `siteID`.

For hzar.doCLTData1DRaw:

If for any locality, there is only a small number of samples taken, warnings will be issued.

If at any locality, the sample variance is 0, a warning is issued, and additional variance is included by estimating the amount of variance ignored due to measurement error.

For hzar.doNormalData1DRaw:

Use the helper function hzar.mapSiteDist to generate site.dist.

The hzar.obsData object created is meant for use with the models constructed using hzar.makeCline1DNormal.

A hzar.obsData object, using the site dinstances and sample means and variances as calculated from the values given.

Graham Derryberry asterion@alum.mit.edu

hzar.obsData For a description of this object structure. hzar.makeCline1DNormal Make models of normal data. hzar.first.fitRequest.gC For compiling those models. hzar.doFit For fitting the compiled models. hzar.doFit.multi For quickly fitting multiple models. hzar.mcmc.bindLL For viewing the mcmc trace.

data(manakinMorphological);
data(manakinLocations);
mknBL <-
  hzar.doNormalData1DRaw(
    hzar.mapSiteDist(siteID=manakinLocations$LocalityID,
                     distance=manakinLocations$distance),
    traitSite=manakinMorphological$Locality,
    traitValue=manakinMorphological$beard.length)
hzar.plot.obsData(mknBL)
mknBLm <-
  hzar.makeCline1DNormal(mknBL, tails="none");
mknBLm <-
  hzar.model.addBoxReq(mknBLm,-30,600);
## A quick way to reduce the number of variables

## Assume the observed means to left and right are the
## population means
hzar.meta.fix(mknBLm)$muL <- TRUE
hzar.meta.fix(mknBLm)$muR <- TRUE

## Make the initial variance to the left and right match the
## observed local variance at the left (site A) and right (site L). 
hzar.meta.init(mknBLm)$varL<-mknBL$frame["A","var"]
hzar.meta.init(mknBLm)$varR<-mknBL$frame["L","var"]

## Assume the observed variance to left and right is the
## the respective population variance. 
hzar.meta.fix(mknBLm)$varL <- TRUE
hzar.meta.fix(mknBLm)$varR <- TRUE

## Now mknBLm has only 3 free parameters instead of 7.

mknBLFR <-
   hzar.first.fitRequest.gC(gModel=mknBLm,
                            obsData=mknBL,
                            verbose=FALSE);
mknBLFR$mcmcParam$chainLength <- 2e3;
mknBLFR$mcmcParam$burnin <- 5e2;
mknBLF <- hzar.doFit(mknBLFR)
plot(hzar.mcmc.bindLL(mknBLF))
## Not run: 
mknBLFR2 <- hzar.next.fitRequest(mknBLF)
## Do more fitting
mknBLFR2$mcmcParam$chainLength <- 1e5;
mknBLFR2$mcmcParam$burnin <- 1e3;
mknBLF2 <- hzar.doFit(mknBLFR2)
plot(hzar.mcmc.bindLL(mknBLF2))

## End(Not run)