Hello,
I went through the idea page for GSoC 2013. I am interested in the working
on the idea of multilayer perceptron implementation. I went through
https://gist.github.com/2061456 posted there. Does this require a lot of
mathematical background to work on (meaning do undergraduates get selected
Hello all,
Using lassoCV to select the optimal regularization I get an extremely high
alpha (1x10^6), which produces all zero weighted features when training the
model.
What I would like to know is if I can infer from such a high parameter that
there is insufficient information contained in the fe
That's right, I forgot to do "make cython" in the base dir. Now the snippet
runs fine. Hopefully it will finish well also with the real data. Thanks a
lot for the great support!
-- Javier Romero --
On Mon, Apr 1, 2013 at 1:06 AM, Doug Coleman wrote:
> You have to run cython on it. Make cython
