[pymvpa] DecMeg2014 data analysis competition

[Emanuele Olivetti]

Hi Yarick,

Answers below,

On 04/23/2014 12:04 AM, Yaroslav Halchenko wrote:
> Hi Emanuele,
>
> Few questions:
>
> 1.  "The goal of this competition is to predict the category of a visual
> stimulus presented to a subject from the concurrent brain activity."
>
> to a NEW (unseen) subject or a subject among those present in training
> set?

The test set is made of new trials from NEW subjects, i.e. subjects not
present in the train set. The underlying idea is the following: when you
want to make inference about the population, you want to be able
to say something about future (new) trials of future (new) subjects.
Of course this is difficult :)
You can have more technical details here:
   http://arxiv.org/abs/1404.4175
(this document is referenced in the competition main page too)

>
> 2.  what is the license on the dataset?  could it be
> redistributed/acquired without signing in with g+/whatever?

The dataset of the competition is released under the Creative Commons
Attribution v3.0 license [*]. At the moment we plan to distribute it through
Kaggle - which, as you point out, require registration or, alternatively, signing
in with some social network account, at your will.
After the competition the dataset will be available from other sources
as well. If you are interested in the dataset now, without entering the
competition, then I guess we can find a solution. We did not anticipate this
scenario, but surely we would be glad to accommodate your request.
Just let me know. Of course you would not get the class labels of the test set :D :D

Best,

Emanuele

[*]: See https://www.kaggle.com/c/decoding-the-human-brain/rules