[pymvpa] [Nipy-devel] [ML-news] New paper on statistical tests

[Francisco Pereira]

Thank you for the pointer, Emanuele! Here's a PDF if you have no
access to PRL (one day the journals will all die, and we'll be better
off for it)

http://www.memswave.angstrom.uu.se/Publications/pdf/p0804.pdf

This may also be of interest

"Interval Estimation for a Binomial Proportion"
http://projecteuclid.org/DPubS/Repository/1.0/Disseminate?view=body&id=pdf_1&handle=euclid.ss/1009213286

as it covers various kinds of interval and provides some startling
results about actual coverage versus the % specified for the interval.

cheers,
Francisco

On Wed, Feb 2, 2011 at 5:37 AM, Emanuele Olivetti
<emanuele at relativita.com> wrote:
> (cross-posting to PyMVPA list)
>
> OK :-) . Happy to hear that mistakes can have positive effects
> from time to time.
> So here is another reference that I find very interesting. It
> belongs a related subtopic:
>
> Cross-validation and bootstrapping are unreliable in small sample
> classification
> by: A. Isaksson, M. Wallman, H. Goransson, M. Gustafsson
> Pattern Recognition Letters, Vol. 29, No. 14. (15 October 2008), pp.
> 1960-1965. doi:10.1016/j.patrec.2008.06.018
> http://dx.doi.org/10.1016/j.patrec.2008.06.018
>
> Comments on this article, especially with respect to neuroimaging
> applications like brain decoding, are warmly welcome.
>
> Emanuele
>
> On 01/31/2011 03:13 PM, Yaroslav Halchenko wrote:
>>
>> no no no -- keep coming ;-) may be even please cross-post to pymvpa
>> mailing list ;)
>>
>> On Mon, 31 Jan 2011, Emanuele Olivetti wrote:
>>>
>>>    Multivariate Statistical Tests for Comparing Classification Algorithms
>>>    Learning and Intelligent OptimizatioN LION5 (Best paper award), Jan
>>>    17-21 2011, Rome Italy,
>>>     (to appear) Lecture Notes in Computer Science, Springer.
>
>
> _______________________________________________
> Pkg-ExpPsy-PyMVPA mailing list
> Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
> http://lists.alioth.debian.org/mailman/listinfo/pkg-exppsy-pymvpa
>