As I know, SVM predicts class label directly without probability
information.
The probability of each class is just an estimation.
(Check section 8 of libsvm paper:
http://www.csie.ntu.edu.tw/~cjlin/papers/libsvm.pdf)
On Fri, May 22, 2015 at 3:03 AM zhenjiang zech xu
wrote:
> Hi all,
>
> I test
ok. I will start working on this.
Please let me know if there are other suggestions.
Thanks,
Chyi-Kwei
On Fri, Aug 22, 2014 at 5:23 AM, Gael Varoquaux
wrote:
> On Tue, Aug 19, 2014 at 11:30:51AM -0400, chyi-kwei yau wrote:
>> So, are we still considering include this model in the futu
/onlineldavb_bsd.tar)
So, are we still considering include this model in the future?
If yes, I can create an issue and start working on this.
Thanks,
Chyi-Kwei
On Sun, Nov 24, 2013 at 12:22 PM, chyi-kwei yau wrote:
> Got it. I will ask him if he can relicense the code to BSD or MIT.
>
>
>
> On Su
Hi Subhabrata,
You can find HMM implementation here:
https://github.com/hmmlearn/hmmlearn
And seqlearn also have Viterbi implementation.
https://github.com/larsmans/seqlearn/blob/master/seqlearn/_decode/viterbi.pyx
Best,
Chyi-Kwei
On Fri, May 16, 2014 at 4:39 PM, Subhabrata Banerjee
wrote:
> D
Got it. I will ask him if he can relicense the code to BSD or MIT.
On Sun, Nov 24, 2013 at 12:06 PM, Gael Varoquaux
wrote:
> On Sun, Nov 24, 2013 at 12:05:07PM -0500, chyi-kwei yau wrote:
>> Sure, that's a possibility. There's already a version of the code in
>> the
Hi guys,
I just got response from Matt Hoffman about relicense his onlineLDA code.
Here is his response:
--
Hi Chyi-Kwei,
Sure, that's a possibility. There's already a version of the code in
the gensim project under the LGPL license—let me know if that's still
to restrictive and I'll
ok. I will email Matt Hoffman about the license thing and reply in this
maling list.
Best,
Chyi-Kwei
On Thu, Oct 31, 2013 at 5:04 AM, Olivier Grisel wrote:
> 2013/10/31 Joel Nothman :
> > I imagine LDA and related models would be welcome additions to
> scikit-learn.
> > However, Hoffman's work
Hi guys,
Recently I am trying to combine LDA(Latent Dirichlet Allocation) with some
clustering methods.
So I modified Matt Hoffman's onlineLDA to sklearn's format and use it with
other sklearn modules.
(I changed the input format to sparse matrix and speed up the e-step with
multiprocessing.)
I t
