fixed in
https://github.com/scikit-learn/scikit-learn/commit/24e962cfe1c348d0c1de95f546b2091fe75a2c06
Alex
On Mon, Jul 27, 2015 at 7:00 PM, panfei wrote:
> the site is:
> http://scikit-learn.org/stable/auto_examples/plot_cv_predict.html
>
>
> The last line of the example code will cause the gen
hi,
> Agreed—this is exactly the type of use case I want to support.
> Pickling won't work here, but using HDF5 like MNE does would
> probably be close to ideal (thanks to Chris Holdgraf for the
> heads-up):
>
> https://github.com/mne-tools/mne-python/blob/master/mne/_hdf5.py
For your info Eric L
FYI it's quite common in brain imaging to use correlation of y_pred
and y_test as metric and not MSE. It's for the same reasons Stylanos
explains so I am inclined to say it could be a valuable contrib.
Alex
On Mon, Sep 7, 2015 at 10:24 PM, Stylianos Kampakis
wrote:
> Hi Andy,
>
> It is because
@mblondel I was talking indeed about Pearson correlation.
Unless proven otherwise with examples, I'd stick to Pearson correlation too.
Alex
--
Monitor Your Dynamic Infrastructure at Any Scale With Datadog!
Get real-time
hi everyone,
just a short email to announce that we are welcoming Tom Dupré la Tour
@TomDLT and Jan Hendrik Metzen @jmetzen as core contributors to the
project. Tom with is work on SAG and NMF, and Jan on Gaussian
Processes and probability calibration have had significant impact on
the project and
hi,
Y has to be a dense array to work.
given the size of your problem it may be a problem.
you can open an issue
Alex
On Thu, Oct 15, 2015 at 8:23 AM, Startup Hire wrote:
> Hi Pypers,
>
> Hope you are doing well.
>
> I am doing multi label classification in which my X and Y are sparse
> matr
which one do you have in mind?
There was a LambdaMART PR at some point.
A
On Sat, Oct 24, 2015 at 6:34 PM, olologin wrote:
> Hello folks, i wonder, why scikit-learn doesn't have ranking algos
> implementation? Is it somehow explained or no one implemented it yet?
>
> Thanks in advance.
>
>
have a look at:
http://fa.bianp.net/blog/2012/learning-to-rank-with-scikit-learn-the-pairwise-transform/
Alex
--
___
Scikit-learn-general mailing list
Scikit-learn-general@list
one API issue that needs to be fixed first is the sample_props fit
parameter to pass query ids with samples.
A
--
___
Scikit-learn-general mailing list
Scikit-learn-general@list
> It's a pity that people who contributed to the release are not listed
> anymore.
here is the fix (without mailmap cleaning)
$git shortlog -s -n 0.16.0..0.17
235 Andreas Mueller
151 Olivier Grisel
60 Alexandre Gramfort
58 Gael Varoquaux
43 Gilles Louppe
39 La
> In addition, users
> cannot take advantage of the LB Keogh lower bound of dynamic time warping,
> which can dramatically speed up the nearest neighbors search.
can you give more details on how you would use this in our NN code?
if passing a dtw callable is not good enough, that can justify an a
> How do you plan to represent variable-length time series? Lists of 1d numpy
> arrays work but would be slow I guess. The ideal representation needs to be
> compatible with grid search and fast.
good point. I was thinking of forcing all time series to have the same length.
or use dtype = object l
I would say prototype it and let's see what it implies on the code.
A
--
Go from Idea to Many App Stores Faster with Intel(R) XDK
Give your users amazing mobile app experiences with Intel(R) XDK.
Use one codebase in this
> Is there any way to specify the random_state of the StratifiedKFold CV
> generator that is used when one uses sklearn.grid_search.GridSearchCV(cv=int
> or none)?
no.
> Or would the best method to ensure replicability between
> GridSearchCV runs simply to make a StratifiedKFold instance myself a
On Sun, Jan 3, 2016 at 6:25 AM, Guoqiang Lan, Mr
wrote:
> But it seem to be not possible to define such a constrained coefficient
> matrix in "sklearn". Am I right?
indeed. You'll need to recode. sklearn lasso only works with in memory
ndarray or sparse matrices.
A
hi,
what is the dtype of your input array?
Alex
--
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transacti
hi Nelson,
your priority now should be to find 2 persons who accept to be mentors.
Alex
On Sat, Mar 19, 2016 at 8:36 PM, Nelson Liu wrote:
> Hi Everyone,
> My name is Nelson Liu, and I’m a current undergrad at the University of
> Washington. I’ve been contributing to scikit-learn since November
hi,
sdca is used internally in liblinear so offered by our logreg estimator.
otherwise it's implemented in lightning :
https://github.com/scikit-learn-contrib/lightning
A
On Sat, Apr 9, 2016 at 8:14 PM, Ahmed SaadAliden
wrote:
> Hi,
>
> I am thinking about adding SDCA "Stochastic dual coordin
> Sorry, ICML is at the same dates as the big brain imaging conference, so
> I will not be able to attend (neither the conference, nor a sprint).
same for me. Surprisingly...
Alex
--
Find and fix application performance
at 11:39 AM, Andreas Mueller wrote:
>>> So should we pick another or possibly an additional date?
>>> Will anyone be in NYC for ICML / UAI / COLT?
>>>
>>> On 04/12/2016 03:56 AM, Alexandre Gramfort wrote:
>>>>> Sorry, ICML is at the same date
hi Will,
> if not sp.issparse(X):
>
> self.reconstruction_err_ = norm(X - np.dot(W, H))
>
> else:
>
> norm2X = np.sum(X.data ** 2) # Ok because X is CSR
>
> normWHT = np.trace(np.dot(np.dot(H.T, np.dot(W.T, W)), H))
>
>
hi Dave
I recommend you start working with the current development
(master branch) version. You'll have to learn a bit of git fu
to create an hmm branch which will allow you to work directly
in hmm.py keeping the old file in the master branch.
See:
http://scikit-learn.org/stable/developers/index
> Er, it looks like safe_sparse_dot() returns sparse unless dense_output=True.
> And, I'm confused as to how this would result in more memory. Aren't we
> allocating more in the lines above for the issparse(X) case? I'm stick right
> now because my 40k x 220k CSR matrix can't make it past computing
I would be in favor of a HierarchicalClustering object that supports
various linkage
criteria.
something like:
hc = HierarchicalClustering(linkage='single')
linkage='ward' would be another option.
Alex
On Sat, Sep 7, 2013 at 4:25 PM, Jacob Vanderplas
wrote:
> On Sat, Sep 7, 2013 at 5:21 AM,
hi James,
for a given value of l1_ratio, the grid of alphas is chosen in log scale
starting from alpha_max to alpha_max / 10**eps. Any value of alpha
larger than alpha_max will lead to a coef_ full of zeros.
HTH
Alex
On Fri, Oct 11, 2013 at 9:39 PM, James Jensen wrote:
> How is the default gri
> I did find the part in coordinate_descent.py where alpha_max is chosen, but
> I don't fully understand the reasoning behind it:
>
> alpha_max = np.abs(Xy).max() / (n_samples * l1_ratio)
it can be derived from the KKT optimality conditions of the Lasso problem.
A
---
> Also is it impacted by the lack of greedy data-centering in the sparse
> case? It seems it does to me.
indeed ! I guess we have a bug there...
I'll take a look later today.
Alex
--
October Webinars: Code for Performan
hi,
here are the commits on SOM from 3 years ago
https://github.com/scampion/scikit-learn/commits/master
Alex
On Tue, Oct 22, 2013 at 8:07 AM, Andreas Mueller
wrote:
> Hi Taylor.
> Thanks for wanting to contribute.
> I am a bit ambivalent wrt to adding SOMs.
>
> I have not seen or heard of an
> By the way, the MiniBatchDictLearning can be trivially modified to do
> this: do a non-negative Lasso, instead of a Lasso. This is discussed in
> the original paper.
if somebody has some time to add a positive option to LassoLars like
available in Lasso that would be great. It would then be stra
> I'd love to add non-negative lasso to this mix. However, I noticed
> that cd_fast.pyx is missing the positive=True option in multitask
> lasso (as well as the sparse variant). Is there any other reason for
> this or just that nobody needed it?
indeed nobody needed it :)
thanks for looking int
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
Gabriel,
your problem is that you want to apply an LDA classifier on a regression task.
Make sure y is discrete (eg. filled with 0 or 1) and it should work.
A
On Sat, Nov 9, 2013 at 7:14 PM, Gabriel Peschl wrote:
> Hi,
>
> I am trying to implement the LDA algorithm using the sklearn, in python
hi Paolo,
I think too that a cross_val_confusion or cross_val_confusion_matrix function
in sklearn.cross_validation would be handy.
Maybe you can contribute yourself? or at least open an issue?
Best,
Alex
On Mon, Nov 11, 2013 at 4:55 PM, Paolo Di Prodi
wrote:
> Hello there,
> correct me if I
Gabriel,
as I already said you're using a classifier for a regression problem.
Please read the doc.
Maybe :
http://scikit-learn.org/stable/tutorial/statistical_inference/supervised_learning.html
this help.
hint : replace LDA by sklearn.linear_model.Ridge for example.
HTH
Alex
On Sat, Nov 9,
hi,
have a look at:
https://github.com/scikit-learn/scikit-learn/pull/2285
any help to test/review this PR is very welcome.
Alex
On Sun, Dec 8, 2013 at 1:53 PM, Chen Wang wrote:
>
> Dear all,
>
> I mainly use scikit learn do regression analysis. I found that this package
> didn't have polyno
> I don't know if Alex will have time to work on it in the near future.
unlikely... sadly...
any help very welcome.
Alex
--
Rapidly troubleshoot problems before they affect your business. Most IT
organizations don't ha
hi,
I feel the same as Jake.
Hadayat I looked at your code and there is a fair bit of refactoring
to be done to fit with the scikit-learn API.
I would encourage you to do this refactoring and try adapting
the current GP examples to see how it compares in terms of
speed, results and code readabil
hi,
did you try SVR ? eventually setting epsilon to 0.?
if it's too slow have a look at lightning new LinearSVR estimator.
Alex
On Sat, Jan 11, 2014 at 7:28 PM, florian.wilh...@gmail.com <
florian.wilh...@gmail.com> wrote:
> Hi,
>
> at Blue Yonder we often use Scikit-Learn but are sometimes
> I like the locality-sensitive hashing idea!
+1
we need to cleanup the GSOC idea wiki page...
Alex
--
WatchGuard Dimension instantly turns raw network data into actionable
security intelligence. It gives you real-time
a hack that might be good enough could be to use as features the cos
and sin of the angle and split the output (using then multioutput
trees)
Alex
--
WatchGuard Dimension instantly turns raw network data into actionable
> Would you be able to tell me which part of the code-base I should play with
> for this project? (I'm assuming it is the cython code in
> coordinate_descent.pyx)
yes.
> Some references to literature would definitely help.
you have a few on the wiki page.
you can also start from wikipedia:
htt
hi,
We are looking for a Scikit-Learn/Python fan interested in helping us
to implement native persistence of Scikit-Learn estimators and data.
The technology we plan to use is called NEO (http://www.neoppod.org/).
It is a distributed object database that can store serialized python
objects on a re
the idea of dropping LibLinear for Logisitic Regression has been
around for some time now. If we manage to have a least the same
performance, supporting both L1 or L1+L2 regularization, without
penalizing the intercept ... it would be great.
@mblondel any thought wrt to lightning ?
Alex
Tom,
for linear regression models (Lasso et al.) there is no support but
you can achieve what you want with manual prewhitening (taking care of
feature scaling c.f. normalize parameter).
Adding native support is conceivable for the project I think.
HTH
Alex
On Tue, Feb 11, 2014 at 3:28 PM, Tho
hi sklearners,
we're planing to reproduce the success of last year's scikit-learn
sprint in Paris.
We'll have a new sprint in Telecom ParisTech in 2014 !
Dates are : 15th-20th of July 2014
(just before EuroPython in Berlin)
I've updated the wiki page:
https://github.com/scikit-learn/scikit-le
hi Clemens,
you might want to look at the open issues on the topic.
A
On Mon, Feb 17, 2014 at 12:52 PM, Clemens Brunner
wrote:
> Hi there!
>
> I would like to use a shrinkage LDA classifier (that is, LDA with Ledoit-Wolf
> covariance estimation). However, I have only found the two necessary
>
https://github.com/scikit-learn/scikit-learn/issues/1649
have a look at the implemention from @ksemb
https://gist.github.com/ksemb/4c3c86c6b62e7a99989b
and the discussion on the issue
HTH
A
--
Managing the Performance
hi Joseph,
yes I would vote for it. More generally probability calibration has been
on the wish list for some time now. See this old PR that needs some
love:
https://github.com/scikit-learn/scikit-learn/pull/1176
any help on this one too is more than welcome.
Best,
Alex
---
Hi,
I like the [MRG+1] and [MRG+2] idea. Let's see if it can help...
Best,
A
--
Flow-based real-time traffic analytics software. Cisco certified tool.
Monitor traffic, SLAs, QoS, Medianet, WAAS etc. with NetFlow Analyze
hi,
do you know:
http://scikit-learn.org/stable/modules/generated/sklearn.pipeline.FeatureUnion.html
?
it might do already what you want
A
On Thu, Feb 27, 2014 at 8:33 AM, michael kneier
wrote:
> Hi all,
>
> I would like to add a "combiner" class which would work with pipeline to
> allow u
hi Manoj,
looks like a pretty decent proposal to me.
Cheers,
Alex
On Thu, Mar 6, 2014 at 6:41 PM, Manoj Kumar
wrote:
> Hello,
>
> I have prepared a wiki page for the first draft of my GSoC proposal after
> several discussions. Please do have a look and provide me feedback.
> https://github.com
> I transform my data using different transformation T1 and T2 and then feed
> it into the LInearSVC. What I found is that the classifier is trained
> significantly faster with the transformed data using T2.
>
> Since both transformed data has the same number of instances, we are looking
> at facto
hi Johannes,
I am personally interested by this code. Can you make it a bit
more userfriendly adding an example as we do with sklearn?
for sklearn, to start, what I recommend is to adapt it to use the sklearn API
and add it to the wiki page of related projects.
best,
Alex
On Fri, Apr 11, 2014
it is used in the get_support method.
https://github.com/scikit-learn/scikit-learn/blob/master/sklearn/linear_model/randomized_l1.py#L121
A
--
Put Bad Developers to Shame
Dominate Development with Jenkins Continuous Inte
hi,
you need to do it manually accessing the .coef_ and .intercept_ attributes
after fitting the model.
HTH
Alex
On Mon, Apr 21, 2014 at 6:58 AM, Michael Kneier
wrote:
> Hi All,
>
> I am using the LogisticRegression model on a sparse matrix of features. Due
> to the certain technical restricti
hi Manoj,
congrats ! Glad to mentor your project this summer.
> One last question, does the PSF or scikit-learn have some sort of convention
> for GSoC, or is it the usual, write -> push -> review -> rewrite -> push ->
> merge cycle?
nothing special for the GSOC. You'll do what you've been doing
hi sklearners,
FYI Danny Sullivan https://github.com/dsullivan7
will work at Telecom ParisTech with me as a scikit-learn
engineer starting this summer. These topics
(SGD improvements, averaging, SAG etc.) are part
of the roadmap.
I think he will start by setting up a benchmark for online
supervis
hi Chris,
you should share a gist on gist.github.com with a .npy containing the
data to reproduce the problem.
Best,
Alex
On Thu, May 15, 2014 at 2:15 AM, Chris Waigl wrote:
> Hi sklearn community,
>
> I'm new on this list, Python user of many years, and maybe an advanced
> beginner with sciki
hi,
fixed in master:
https://github.com/scikit-learn/scikit-learn/commit/4a04b3fb076232542c4f73683bd8ea86e2ce3c81
thanks for reporting the issue
A
On Tue, May 20, 2014 at 9:16 AM, Peiyun Hu wrote:
> Hello everyone,
>
> I am currently using the Dictionary Learning of Scikit-Learn 0.14. As
> t
hi,
the gram computation is correct.
what I propose is to change :
regularization = max(n_features / 10, 1)
by
regularization = max(n_components / 10, 1)
or
regularization = min(max(n_features / 10, 1), n_components)
wdyt?
Best,
Alex
---
see
https://github.com/scikit-learn/scikit-learn/pull/3176
Alex
--
"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
Instantly run your Selenium tests across 300+ browser/OS combos.
Get unparalleled
can you send a PR and use github to report
issues? it's easier to track compared to an email...
thanks
A
--
"Accelerate Dev Cycles with Automated Cross-Browser Testing - For FREE
Instantly run your Selenium tests across 3
hi,
I don't get the problem with this docstring.
it would be great to merge:
https://github.com/scikit-learn/scikit-learn/pull/3176
Alex
--
Time is money. Stop wasting it! Get your web API in 5 minutes.
www.restlet.com
> "
> "of features")
>
> is making sure the number of n_nonzero_coeffs is less than the number of
> atoms (which seems to be the limiting factor in using the Gram matrix), not
> the number of features. Should this error string also be
hi,
can you open an issue on github and
provide a gist to reproduce the pb?
thanks
Alex
On Fri, Jun 6, 2014 at 12:16 PM, Jiancheng Sun wrote:
> Got an error "AttributeError: 'MiniBatchDictionaryLearning' object has no
> attribute 'random_state_'" when I use
> MiniBatchDictionaryLearning.partia
On Thu, Feb 13, 2014 at 8:04 PM, Alexandre Gramfort
wrote:
> hi sklearners,
>
> we're planing to reproduce the success of last year's scikit-learn
> sprint in Paris.
>
> We'll have a new sprint in Telecom ParisTech in 2014 !
>
> Dates are : 15th-20th of Jul
hi Brett,
your code looks good.
can you share a full gist showing clearly the problem?
my first check would be to compare using X.toarray()
to make sure the dense / sparse code agree.
Alex
On Mon, Jun 9, 2014 at 9:36 PM, Brett Meyer wrote:
> I’m having an issue using the prediction probabil
e 1k-2k samples that all have
> the same prediction value, and there are only 3600 distinct prediction
> values over all of the folds for cross-validation. The resulting ROC
> looks like five big stair steps, with some little bits of fuzziness around
> the inner corners.
>
>
>
>
for the record here is the issue:
https://github.com/scikit-learn/scikit-learn/issues/3266
Alex
--
HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HP
Fri, Jun 13, 2014 at 3:42 PM, Nicolas Goix wrote:
> Hello,
>
> This is my first post to the list, I have been recently in touch with
> Alexandre Gramfort, and I would be very interested in exploring some
> outlier/anomaly detection algorithms, before eventually put it in a
> compa
hi,
> I was wondering if there is any reason of why the randomized l1 algorithm
> from the stability selection paper is implemented only using Lars Lasso and
> not the coordinate descent algorithm.
> I think than including a version of the algorithm with the coordinate
> descent method would be ve
hi,
have you played with the sparseness parameter of the NMF estimator?
Alex
On Wed, Jun 25, 2014 at 3:42 PM, Vamsi Krishna Potluru
wrote:
> Hello,
>
> I have worked a bit on the sparse NMF model proposed by Hoyer [1]. The
> paper is mentioned in the Scikits NMF module but AFAIK the model is
hi,
proba calibration with libsvm (using Platt's method) involves data resampling.
So between runs the result can change.
HTH
Alex
On Thu, Jun 26, 2014 at 12:51 AM, Stelios wrote:
> Hello all,
>
> I have the following code:
>
> . . . .
> # 'train' is a (M,N) numpy array (input) and 'traint' is
hi,
> sorry I did not understand. Are you asking me to change the code in order to
> add these features?
if you need it and get better/faster results with it please go ahead.
add a parameter
method='lars' | 'cd'
We'll discuss on the pull-request
Alex
> I can probably rewrite the randomized l
hi,
I would be +1 adding an invert_whitening param to PCA that would
default to False in 0.15 and move to True in 0.16 to eventually
disappear later.
Alex
On Mon, Jun 30, 2014 at 8:53 AM, Michael Eickenberg
wrote:
> Kyle is facing the same question for his incremental pca pr
> https://github.c
hi,
proper PCA is run on centered data (axis=0) otherwise it's a truncated SVD.
I seams you want a PCA on X.T (X transposed).
HTH
Alex
On Tue, Jul 22, 2014 at 3:14 AM, Adam Hughes wrote:
> Hi,
>
> I'm really enjoying scikit learn and looking to add a lite version of PCA to
> some programs I'm
> Is it possible to use scipy sparse matrices with LinearSVC?
sklearn.svm.LinearSVC will accept sparse data.
it's the same classes that should work with both dense and sparse data
A
--
Want fast and easy access to all t
> So how do I obtain the class probability along with classification?
you help me finish :
https://github.com/scikit-learn/scikit-learn/pull/1176
:)
Alex
--
Want fast and easy access to all the code in your enterprise?
> But now it makes me think -
> How OneVsRestClassifier approach is different then SGDClassifier?
> Is SGDClassifier an optimization algorithm which also uses
> OneVsRestClassifier for classification?
yes SGDClassifier uses OvR internally.
A
--
> But SGDClassifier optimizes classification-specific loss functions,
> unlike ElasticNet which is a regressor.
indeed but squared loss is cheap to use and can reach pretty good
classif performance in practice.
A
--
Want
hi Laszlo,
thanks for using the bug tracker and not the ML to report issues.
Also avoid pasting raw text/code in emails but use services like gist.github.com
what version of sklearn are you using? ElasticNetCV uses threads in v0.15
so the memory problem should not happen anymore.
Alex
On Fri,
hi Neal,
you should concat the imaginery and real parts of the features.
X = np.c_[X.real, X.imag]
if you use the euclidian distance it should do the join.
Alex
On Wed, Sep 17, 2014 at 8:05 PM, Mohamed-Rafik Bouguelia
wrote:
> Hi,
> You cannot use complex numbers, they should be real number
hi,
> I am trying to make the K-SVD implementation consistent with the existing
> sparse coding algorigthms (dict_learning and orthogonal_mp) and I am a bit
> confused by the notations. I'll give a summary of the notations and explain
> what I am not understanding.
>
> 1) sklearn.decomposition.dic
Indeed I can reproduce the problem.
Maybe @MechCoder can have a look.
out of curiosity why don't you use RidgeCV?
Alex
--
Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer
Achieve PCI DSS 3.0 Compliant Sta
chris please open an issue.
it also seems like the issue appears with lsqr but not with other solvers.
Alex
--
Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer
Achieve PCI DSS 3.0 Compliant Status with Out
hi Guillaume,
have a look at:
http://scikit-learn.org/stable/developers/index.html
HTH
Alex
On Thu, Nov 13, 2014 at 10:22 AM, Guillaume Dumas
wrote:
> Hi!
>
> I am trying to implement a recent methods of non-negative matrix
> factorization for scikit-learn.
> To date, I adapted the existing s
hi,
in this branch you have a toy implementation that does not scale
https://github.com/agramfort/scikit-learn/tree/mkl
Maybe it helps,
Alex
On Wed, Nov 26, 2014 at 9:23 PM, Andy wrote:
> Hi Arman.
> There is no multiple kernel learning that automatically learns the weights
> of the different
> Do you have some examples for this?
just read the commit:
https://github.com/agramfort/scikit-learn/commit/34223251fa33ae25c80f8e1dfb630a2e1c5ba7c6
A
--
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Serve
What I suggest:
use isinstance(estimator, RegressorMixin) to know if we can use predict safely.
If we can't rely on inheritance, call predict if neither predict_proba
nor decision_function are available and check that the predicted
values are of type float32 or float64.
Alex
FYI Mathieu and myself converged on this with Danny, believing it's
the easiest move.
It could later be improved and extended with schemes that also use feature
specific learning rates.
any thought?
Alex
On Tue, Dec 2, 2014 at 2:48 PM, Daniel Sullivan wrote:
> Hey All,
>
> I've been looking a
> A compromise would be to just implement the Cython routine in a separate
> file, while sharing the same file for the pure Python side.
sounds reasonable.
> That said, using a separate class for Adagrad would allow to get rid of
> irrelevant hyper-parameters.
+1
> Some code from the SGD module
there is an "allow_nd" param I added to support this use case. Maybe
we could add it to GridSearchCV...
Alex
--
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Bu
> I just looked at the list from last year, and what seems most relevant
> still is GMMs,
> and possibly the coordinate descent solvers (Alex maybe you can say what
> is left there or
> if with the SAG we are happy now?)
there is work coming in coordinate descent and SAG is almost done.
I don't th
FYI I created the wiki page but it needs editing. So it's WIP
https://github.com/scikit-learn/scikit-learn/wiki/Google-summer-of-code-(GSOC)-2015
A
--
Dive into the World of Parallel Programming. The Go Parallel Website,
> I could catch the warning but this is probably a bad idea? What is the
> new, preferred approach to "fit_intercept=True"?
the way to go is to *not* use fit_intercept in lasso_path and do it outside.
we're going to remove this option.
Alex
--
yes or use Lasso directly.
A
--
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software devel
please wait a bit so we finalize the list. It's not definitive.
A
On Mon, Feb 9, 2015 at 6:39 PM, ragv ragv wrote:
> Hi,
>
> I saw implementing GAMs as one of the suggested topics for GSoC 2015.
> Could I take that up? I saw your ( Alex's ) name under that. If yes,
> please let me know I'll star
fit_intercept is not deprecated in Lasso* and ElasticNet* classes. It
defaults to True.
A
--
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slash
what do you have in mind?
A
--
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software develo
my short list is:
GMM
GP
PLS/CCA
so consolidate what we have.
Alex
--
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hu
1 - 100 of 370 matches
Mail list logo
