Hi Andrew,
>>
>> Does this sidestep the degrees of
>> freedom problem discussed by Garland et al.? Can anybody point me to
>> references discussing the mechanics of this process and why this is an
>> appropriate thing to do?
>
Others on this list will disagree with me, but it's not a "degree
Hi Emmanuel,
Great point that it will be faster to simulate from the root up rather than
compute the matrix for large trees. In that case, how about simulate from
the equation
x(t'') = x(t') e^{-alpha (t''-t') }+theta (1-e^{-alpha (t''-t') }) + sqrt(
sigma^2( 1 -e^{-2\alpha (t''-t') )/2 alpha )
Ah, so while re-creating my problem for copy-paste-debug
goodness on the listserv, I discovered what was confusing me.
Originally, when I ran the various models, I got these
log-likelihoods for results:
==
tf2ic2kzkr t
Hi Nick-
Are you are getting differences in relative AICs between models from simple
rescaling (multiplying by a constant)?
The actual values of the traits *might* matter for optimization, depending on
various parameters associated with optimization (and whatever algorithm is
being used - th
Doh! Really should have remembered that,
likelihoods-can-be-greater-than-1 is likelihood 101...
I am still a little puzzled by the dramatically different
results between rescaling and not, will try to post an
example in a sec...
On 3/7/11 12:37 PM, Nick Matzke wrote:
Hi all,
It seems to
Hi Rob,
this is partly also possible in R. You can find all generic functions
specificly written for an (S3) object using:
methods(class="phylo")
This will miss functions, which are not generic and generic function
which use the default mechanism, e.g. class "pml" has a specific
"logLik.pml" meth
Nick,
Log-likelihoods are calculated as the logarithm of the product of the
heights of the probability density function. Since the probability
density function must integrate to 1.0, it can have a height that is
much greater than 1.0 if all the probability density is concentrated on
a small
Hi all,
It seems to be a popular week for questions!
I am running fitContinuous on a variety of continuous trait
data. I am noticing that when the traits are in units where
the max is less than 1 (these are not ratio data, though),
many of the various models produce log-likelihoods that are
Hi All,
On the issue of visibility of functions in R, I definitely find it a
struggle to 'discover' things that I could do with a particular object.
Python has this worked out in a really nice way - if I have an object I can
simply put a full-stop after it and hit tab, and I get a list of all
curr
Hi Andrew,
> As I understand it, gls() is doing a multiple generalized LS
> regression with as many dummy variables as there are factor levels.
> Is this a correct characterization?
I think you'd get one dummy variable less than factor levels in your
characterization (at least in regards to the
Hi everyone,
I am trying to piece together the current best-practices for
"phylogenetic ANOVA" with multi-state predictors.
In my dataset, my four-level factor is non-random with respect to
phylogeny. That is, if I know which higher level clade an species
belongs to, I can predict with pretty go
Emmanuel, all-
That's pretty fast, even on my cheap laptop!
> system.time(desc1<-c(as.list(1:Ntip(res_tree)),prop.part(res_tree)))
user system elapsed
0.650.000.66
As far as why I didn't try prop.part, to be honest, I had no idea that
prop.part() did that. The help file says:
Desc
Hi Alanna,
It looks like the alpha of your fitted OU model is very large (>800)?
In this case the corStruct from corMartins will essentially be a
diagonal matrix of ones and the estimated ancestral state using GLS will
be the arithmetic mean of your samples.
Thus, your result makes sense - a
Hi Alanna -
I can't vouch for the r code but if you have an OU model with a very high
alpha, then you should expect all of the ancestral state estimates to be
exactly the "optimal" trait value under OU - which will be the mean of all the
tip species.
Luke
On Mar 7, 2011, at 3:59 PM, Alanna
Dear all
I am trying to estimate ancestral characters using 'ace' in ape. I'm
using GLS with a corStruct based on an alpha I estimated using OUCH, and
corMartins. The estimate I get for the root node for each of my
parameters exactly matches the mean of the tip values (to at least 6
decimal places
Dear r-sig-phylo mailing list,
The problem with reading a pre-computed species tree from
www.microbesonline.org has been solved. See the text for details.
Kind regards,
Thierry Janssens
-Original Message-
From: Emmanuel Paradis [mailto:emmanuel.para...@ird.fr]
Sent: maandag 7 maart 20
16 matches
Mail list logo