Re: [Scikit-learn-general] Self Organizing Map implementation

2013-10-21 Thread Andreas Mueller
Hey Taylor. Currently I'd say +1 and add it to the neural networks folder as a reference implementation. Not sure what the other think. For the visualization, there are some examples of 2d visualization where I guess the SOM would be a natural comparison: http://scikit-learn.org/dev/auto_exampl

Re: [Scikit-learn-general] Self Organizing Map implementation

2013-10-21 Thread Taylor Sather
Hey Andy, Haha, I really appreciate the quick and candid response! In all honesty, I liked them less and less as I applied it to problems in the wild :) I was intrigued by the idea of extremely large grids (> 2k neurons) and the "emergent" behavior that the som would exhibit under theses ci

Re: [Scikit-learn-general] Self Organizing Map implementation

2013-10-21 Thread Alexandre Gramfort
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

Re: [Scikit-learn-general] Self Organizing Map implementation

2013-10-21 Thread Andreas Mueller
Hi Taylor. Thanks for wanting to contribute. I am a bit ambivalent wrt to adding SOMs. I have not seen or heard of an application where SOMs work better than any of the clustering or manifold-learning algorithms in sklearn. On the other hand, it is a classical algorithm and having a reference im

[Scikit-learn-general] Self Organizing Map implementation

2013-10-21 Thread Taylor Sather
Hello, I was wondering if there was any effort to implement a Kohonen map in scikit-learn? I'm thinking of getting my implementation up to snuff for a pull request, but I wanted to ask the mailing list before I invested too much effort. Thanks, Taylor Sather --

Re: [Scikit-learn-general] Double NMF

2013-10-21 Thread Andreas Mueller
On 10/21/2013 02:30 PM, Olivier Grisel wrote: > 2013/10/21 : >> Dear Sklearn Team, >> >> I want to fork the ProjectedGradientNMF -> DoubleProjectedGradientNMF code >> with a minor change that I hope will have useful applications. >> >> I want the code to work on two matrices M1, M2, which are alig

Re: [Scikit-learn-general] Double NMF

2013-10-21 Thread Olivier Grisel
2013/10/21 : > Dear Sklearn Team, > > I want to fork the ProjectedGradientNMF -> DoubleProjectedGradientNMF code > with a minor change that I hope will have useful applications. > > I want the code to work on two matrices M1, M2, which are aligned in the > observations (rows), but have different f

[Scikit-learn-general] Double NMF

2013-10-21 Thread desitter . gravity
Dear Sklearn Team, I want to fork the ProjectedGradientNMF -> DoubleProjectedGradientNMF code with a minor change that I hope will have useful applications. I want the code to work on two matrices M1, M2, which are aligned in the observations (rows), but have different features. Then an observati

Re: [Scikit-learn-general] LabelBinarizer for large data

2013-10-21 Thread Mahendra Kariya
I have added a comment on the pull request.   Regards, Mahendra Kariya On Monday, 21 October 2013 5:53 PM, Arnaud Joly wrote: The main feature is implemented, but there is still some works >to ensure that it works correctly in all edge case. Furthermore, some >modification have been suggested

Re: [Scikit-learn-general] LabelBinarizer for large data

2013-10-21 Thread Arnaud Joly
The main feature is implemented, but there is still some works to ensure that it works correctly in all edge case. Furthermore, some modification have been suggested by Joel. If you want to finish the work of Rohit Sivaprasad, you can ask him in the pull request. Best regards, Arnaud On 21 Oct

Re: [Scikit-learn-general] LabelBinarizer for large data

2013-10-21 Thread Olivier Grisel
2013/10/21 Mahendra Kariya : > Ohh Thanks a lot Arnaud! > > I was actually going to implement sparse matrix for this issue, but it is > already done in this pull request. > BTW when will this change be merged to the master branch? > > On a side note, what IDE do you guys prefer for python? I was ju

Re: [Scikit-learn-general] Hidden Markov Model for one dimensional time series data

2013-10-21 Thread Alexandr M
No, it works, thanks! I used model.fit(signal) instead of model.fit( [signal] ) model = GaussianHMM(n_components = 2) s1 = np.random.randn(50,1) s2 = np.random.randn(50,1)+5 signal = np.concatenate([s1, s2]) model.fit([signal]) BR, Alexandr On 21 October 2013 01:42, Robert McGibbon wrote: >

Re: [Scikit-learn-general] Hidden Markov Model for one dimensional time series data

2013-10-21 Thread Alexandr M
Hi Robert, No, it doesn't work: model = GaussianHMM(n_components = 2) s1 = np.random.randn(50,1) s2 = np.random.randn(50,1)+5 signal = np.concatenate([s1, s2]) model.fit(signal) .../lib/python2.7/site-packages/sklearn/hmm.pyc in _init(self, obs, params)754 self.n_features))755 --> 756

Re: [Scikit-learn-general] LabelBinarizer for large data

2013-10-21 Thread Mahendra Kariya
Ohh Thanks a lot Arnaud! I was actually going to implement sparse matrix for this issue, but it is already done in this pull request. BTW when will this change be merged to the master branch? On a side note, what IDE do you guys prefer for python? I was just browsing through the code in master

Re: [Scikit-learn-general] LabelBinarizer for large data

2013-10-21 Thread Arnaud Joly
It sounds like you haven't enough memory to store a dense matrix of binarized labels. There is already one pr that tries to alleviate this problem : see https://github.com/scikit-learn/scikit-learn/pull/2458 Best, Arnaud On 20 Oct 2013, at 20:20, Olivier Grisel wrote: > 2013/10/20 Mahendra