-------- Original Message --------
Subject: Re: discriminant function in MorphoJ
Date: Thu, 7 Jul 2011 12:57:07 -0400
From: Fabio de A. Machado <[email protected]>
To: [email protected]
Dear all,
Is there any work addressing the influence of using observations
classified through DF in the estimation of pooled parameters such as
var-covar matrices? My question is if doing so can in fact improve the
estimations.
best,
Fabio A. Machado
Em 07/07/2011, às 12:57, morphmet escreveu:
-------- Original Message --------
Subject: Re: discriminant function in MorphoJ
Date: Thu, 7 Jul 2011 03:01:25 -0400
From: andrea cardini <[email protected]>
To: [email protected]
Daniela, it has to be 'good' after the cross-validation. A leave-one-out
x-validation in practice simulates the estimate of unknowns and mitigates
issues of overfitting which tend to inflate predictive accuracy.
Also, when you do predictions, give a look at typicality probabilities
(which you won't get from MorphoJ, I think, but can find in other programs).
There's many papers, book chapters etc. on DA. These are just a few
suggested introductory readings:
Albrecht, G. 1992. Assessing the affinities of fossils using canonical
variates and generalized distances, J. Hum. Evol. 7, 4969. (With an
excellent explanation of typicality probabilities).
Kovarovic K., Aiello L.C., Cardini A., Lockwood C.A. Discriminant function
analyses in archaeology: are classification rates too good to be true?
Journal of Archaeological Science, in press. (Mostly on the importance of
x-validations; PREPRINT available upon request).
Huberty, C.J., Hussein, M.H., 2003. Some problems in reporting use of
discriminant analyses. J Exp Educ 71, 177-191. (Distinction between
predictive and descriptive DAs and more).
Strauss, R.E., 2010. Discriminating groups of organisms, in: Ashraf, E.
(Ed.), Morphometrics for Nonmorphometricians. Springer-Verlag Publishers,
Berlin, pp. 73-91. (Another excellent intro that covers a variety of topics).
The book chapter by Neff and Marcus (1980, if I remember well) was also a
great introduction on the subject but it may be difficult to get. I might
have bought one of the last copies available on the web. I hope one day
someone makes a pdf of that book as it is still one of the best readings
for people doing multivariate morphometrics.
Finally, Klingenberg & Monteiro (Syst. Biol. 54(4):678688, 2005) as well
as Mitteroecker & Bookstein (Evol Biol (2011) 38:100114) make some very
interesting points about DAs. These two papers focus on specific aspects of
DAs (and also other methods in K&M) and are not simple introductions, but I
would recommend you to read them after the more introductory stuff.
Good luck.
Cheers
Andrea
At 12:02 06/07/2011 -0400, you wrote:
-------- Original Message --------
Subject: discriminant function in MorphoJ
Date: Tue, 5 Jul 2011 15:57:46 -0400
From: Daniela Sanfelice <[email protected]>
To: [email protected]
Hello...
by the way, yet using MorphoJ...
If I have a "good" discriminant function result for sex, can I try to
apply it to individuals that I don´t know the sex to "estimate" sex in it?
Thank you very much, cheers, Daniela Sanfelice.
Dr. Andrea Cardini
Researcher in Animal Biology
Dipartimento di Biologia, Universitá di Modena e Reggio Emilia, via Campi
213, 41100, Modena, Italy
tel: 0039 059 2055017 ; fax: 0039 059 2055548
Honorary Fellow
Functional Morphology and Evolution Unit, Hull York Medical School
University of Hull, Cottingham Road, Hull, HU6 7RX, UK
University of York, Heslington, York YO10 5DD, UK
Adjunct Associate Professor
Centre for Forensic Science , The University of Western Australia
35 Stirling Highway, Crawley WA 6009, Australia
E-mail address: [email protected], [email protected],
[email protected]
Webpage: http://sites.google.com/site/hymsfme/drandreacardini
Datasets:
http://ads.ahds.ac.uk/catalogue/archive/cerco_lt_2007/overview.cfm#metadata
