2013/11/7 Mathieu Blondel math...@mblondel.org:
On Fri, Nov 8, 2013 at 12:28 AM, Vlad Niculae zephy...@gmail.com wrote:
I feel like this would go against explicit is better than implicit,
but without it grid search would indeed be awkward. Maybe:
if self.alpha_coef == 'same':
About the LBFGS-B residuals (non-)issue I was probably confused by the
overlapping on the plot and mis-interpreted the location of the PG-l1
and PG-l2 curves.
--
Olivier
--
November Webinars for C, C++, Fortran
Re: the discussion we had at PyCon.fr, I noticed that the internal
elastic net coordinate descent functions are parametrized with
`l1_reg` and `l2_reg`, but the exposed classes and functions have
`alpha` and `l1_ratio`. Only yesterday there was somebody on IRC who
couldn't match Ridge with
And lambda is a reserved keyword in Python ;-)
On Fri, Nov 8, 2013 at 4:59 PM, Olivier Grisel olivier.gri...@ensta.orgwrote:
2013/11/7 Mathieu Blondel math...@mblondel.org:
On Fri, Nov 8, 2013 at 12:28 AM, Vlad Niculae zephy...@gmail.com
wrote:
I feel like this would go against
Just my 0.02$ as a user: I was also a confused/put-off by `alpha` and
`l1_ratio` when I first explored SGDClassifier, I found those names to
be pretty inconsistent --- plus I tend to call my regularization
parameters `lambda` and use `alpha` for learning rates. I'm sure other
people associate
SGDClassifier adopted the parameter names of ElasticNet (which has been
around in sklearn for longer) for consistency reasons.
I agree that we should strive for concise and intuitive parameter names
such as ``l1_ratio``.
Naming in sklearn is actually quite unfortunate since the popular R package
We cannot use lambda as parameter name because it is a reserved
keyword of the python language (for defining anonymous functions).
This is why used alpha instead of lambda for the ElasticNet /
Lasso model initially and then this notation was reused in more
recently implemented estimators such as
just a remark in LogisticRegression you can use L1 and L2 reg and
there is a single param that is alpha.
It's not trivial to have a consistent naming for regularization param.
In SVC it is C as it's the common
naming... but it corresponds to 1/l2_reg with what you suggest...
Alex
On Fri, Nov 08, 2013 at 11:56:24AM +0100, Olivier Grisel wrote:
In retrospect I would have prefered it named something explicit like
regularization or l2_reg instead of alpha.
Agreed.
Still I like the (alpha, l1_ratio) parameterization better over the
(l2_reg, l1_reg) parameter set
Did anyone work on this problem (exceptions raised by classifiers in
grid search) since? I would be happy to do some work to fix this
problem, but would need some advice.
It seems to me like the easiest way around the issue is to wrap the call
to clf.fit() in a try statement and catch the
Hi All,
I've added some new ECOC some time ago. Would it be possible to have some
review and feedback? Also, would you recommend any datasets that could be
used for verification? I am especially concerned about what type and size
of data sets I should use.
I would appreciate any help and
Hi All,
Has there any been discussion on adding some automated benchmarks for both
speed and accuracy of the algorithms we have? I think it would very
interesting if such a script could be automatically executed after every
commit so that we could follow the performance of scikit-learn or, at
On Fri, Nov 8, 2013 at 6:30 PM, Karol Pysniak kpysn...@gmail.com wrote:
Hi All,
Has there any been discussion on adding some automated benchmarks for both
speed and accuracy of the algorithms we have? I think it would very
interesting if such a script could be automatically executed after
We have an instance of vbench continuously running [1] that I did as a
GSoC project last year.
For some reason it seems that the links don't generate properly now,
but it still works (though all data got lost in a jenkins setup
incident this summer).
Here are some linear model benchmarks for
Looks good, but I was more interested if we want to have a single script or
set of scripts that would produce a single number that could used to
compare changes. What do you think?
Thanks,
Karol
2013/11/8 Skipper Seabold jsseab...@gmail.com
On Fri, Nov 8, 2013 at 6:30 PM, Karol Pysniak
It should be written in such a way so that you can add more benchmarks
with a PR to that repo (the master branch) and it should just work.
Many parts of the framework are still hackish though.
Yours,
Vlad
On Fri, Nov 8, 2013 at 7:53 PM, Karol Pysniak kpysn...@gmail.com wrote:
Awesome, thanks
Yes, it would be perfect. Do you have any work planned or schedule on this
problem? Let me know if there is something I could help with.
Thanks,
Karol
2013/11/8 Vlad Niculae zephy...@gmail.com
It should be written in such a way so that you can add more benchmarks
with a PR to that repo (the
17 matches
Mail list logo
