The link between BM and WN is even closer than that: WN is the
derivative of a BM process. Now, BM is nowhere differentiable, so in the
usual sense, WN doesn't really exist. However, it can be approximated by
simulation.
Cheers,
Simon.
On 01/02/11 03:53, Luke Harmon wrote:
I agree with Dav
Having the null hypothesis on the extreme of the range of possible
values is a problem, but can be overcome in some situations. E.g. It is
well known that in variance component models, where the null hypothesis
is var=0 for a single variance component, the ordinary LR test is
conservative becau
Ted said --
> One point for clarification and your further thoughts.
> The way parameterize the OU process in Lavin et al. (2008) it is a
> value of zero (not infinity) that gives a start phylogeny with
> contemporaneous tips. Sometimes the ML estimate of d (what we call
> it) goes to zero, but
often it may go very small but not zero.
In either case, it seems to me you can do a LRT versus a star with one
d.f.
Cheers,
Ted
Original message
Date: Mon, 31 Jan 2011 13:02:38 -0800
From: Joe Felsenstein
Subject: Re: [R-sig-phylo] Model-Selection vs. Finding Models that
David Bapst and Cecile Ane noted that
> On Mon, Jan 31, 2011 at 12:45 PM, Cecile Ane wrote:
> > I don't find the white noise to be any good evolutionary scenario: it's
> > nowhere continuous. It just reduces to the assumption of normal,
> independent
> > observations at the tips. Nothing fancy,
Hello All-
On Mon, Jan 31, 2011 at 11:53 AM, Luke Harmon wrote:
> I agree with Dave here. White noise has two parameters, mean and variance,
> and - to me - is an interesting model to test. But I'm not sure it should be
> considered as a "baseline." One can link Brownian motion and white noise
>
m: "Liam J. Revell"
Subject: Re: [R-sig-phylo] Model-Selection vs. Finding Models that
"Fit Well"
To: David Bapst
Cc: r-sig-phylo@r-project.org
>To the original post, what I think Dave might actually want to do
here
>is fit some kind of no
Hi Liam et al.,
Good point. How many parameters does such a "free model" have, and what are
they?
Cheers,
Ted
Original message
Date: Mon, 31 Jan 2011 14:23:15 -0500
From: "Liam J. Revell"
Subject: Re: [R-sig-phylo] Model-Selection vs. Finding
To the original post, what I think Dave might actually want to do here
is fit some kind of no-common-mechanism model in which the evolutionary
process is totally free to vary across all the branches of the tree.
[Luke Harmon pointed out to me that one example of this would be the
'free model' o
I don't find the white noise to be any good evolutionary scenario: it's
nowhere continuous. It just reduces to the assumption of normal,
independent observations at the tips. Nothing fancy, then :)
Cecile.
On 01/31/11 11:53, Luke Harmon wrote:
I agree with Dave here. White noise has two parame
tions:
> http://www.biology.ucr.edu/people/faculty/Garland/GarlandPublications.html
>
> Garland and Rose, 2009
> http://www.ucpress.edu/books/pages/10604.php
>
>
> ---- Original message
>
> Date: Mon, 31 Jan 2011 09:53:25 -0800
>From: Luke Harmon
>Sub
le/faculty/Garland/GarlandPublications.html
Garland and Rose, 2009
http://www.ucpress.edu/books/pages/10604.php
Original message
Date: Mon, 31 Jan 2011 09:53:25 -0800
From: Luke Harmon
Subject: Re: [R-sig-phylo] Model-Selection vs. Finding Models that
"Fit Well"
I agree with Dave here. White noise has two parameters, mean and variance, and
- to me - is an interesting model to test. But I'm not sure it should be
considered as a "baseline."
One can link Brownian motion and white noise through the Ornstein-Uhlenbeck
model - BM is OU with alpha (constraint
Florian-
Doesn't white noise have two parameters, mean and variance, and thus is just
as complex as the Brownian Motion model? I guess LRT could be done against
both.
That said, it isn't clear to me that what it means for the White Noise to
fit best, as WN is interpretable as an evolutionary scena
Hi David and list,
just a quick comment on one of your questions :
for quantitative traits on a phylogeny you can compare your "best" model to
the "white noise" model implemented in geiger, which assumes that your
traits are drawn from a normal distribution.
This last model would be the "baseline
Hello all,
Apologies for leaving the replies to get cold for a week, but now I
finally have some time to respond.
On Thu, Jan 20, 2011 at 12:17 PM, Brian O'Meara wrote:
> I think considering model adequacy is something that would be useful to do
> and is not done much now. One general way to do
If one is interested in absolute goodness of fit, rather
than model comparison (which model fits best, which might
not be useful if you are worried that all your models are
horrible), wouldn't cross-validation be a good technique?
I.e. leave out one tip, calculate the model and the
estimated n
Hi David, List,
I think you make a good point. After all, the goal isn't to match the
pattern but to match the process. If we just wanted to match the data we'd
use the most complicated model we could make (or some machine learning
pattern) and dispense with AIC.
If a model has errors that are
:
http://www.biology.ucr.edu/people/faculty/Garland/GarlandPublications.html
Garland and Rose, 2009
http://www.ucpress.edu/books/pages/10604.php
Original message
Date: Thu, 20 Jan 2011 11:27:37 -0600
From: David Bapst
Subject: [R-sig-phylo] Model-Selection vs. Findi
I think considering model adequacy is something that would be useful to do
and is not done much now. One general way to do this is to simulate under
your chosen model and see if the real data "look" very different from the
simulated data. For example, I might try a one rate vs. a two rate Brownian
Hello all,
I'd like to pose a question to this group, as a bit of topical
discussion. I apologize in advance if I should mangle a concept.
In many model-based PCMs and some other analyses (such as paleoTS), we
fit models to data by finding the ML estimates of the parameters
associated, calculate t
21 matches
Mail list logo
