Peter - did you get any updates from Kaggle? If not, is there anything
that we as a community can do to sway them?
On Sat, Jul 7, 2012 at 7:46 PM, Emanuele Olivetti
wrote:
> On 07/05/2012 04:37 PM, Olivier Grisel wrote:
> > 2012/7/5 Emanuele Olivetti :
> >> On 07/05/2012 08:49 AM, Olivier Grisel
2012/7/9 Alexandre Gramfort :
> hi all,
>
>> In addition to Platt scaling, there is isotonic regression.
>
> I started to play with the calibration based on isotonic regression
> this morning:
>
> https://github.com/agramfort/scikit-learn/commit/5b1017df6ca69c7733f9e5c0c26926ce6c1c8775
>
> it's bas
hi all,
> In addition to Platt scaling, there is isotonic regression.
I started to play with the calibration based on isotonic regression
this morning:
https://github.com/agramfort/scikit-learn/commit/5b1017df6ca69c7733f9e5c0c26926ce6c1c8775
it's based on mathieu's calibration_plot PR
a sklear
In addition to Platt scaling, there is isotonic regression. Two references:
- Transforming Classifier Scores into Accurate Multiclass Probability
Estimates (KDD 2002)
- Predicting accurate probabilities with a ranking loss (ICML 2012)
Mathieu
On Mon, Jul 9, 2012 at 10:53 PM, Mathieu Blondel wrote
> To be honest, I haven't worked with Platt scaling yet - Paolo has
> discussed it on the mailing list a while ago [1] - maybe he has some
> code to share.
>
Paolo's code is here:
https://github.com/paolo-losi/scikit-learn/tree/calibration
Mathieu
-
On 07/09/2012 01:56 PM, Peter Prettenhofer wrote:
> 2012/7/9 Emanuele Olivetti :
>> Hi Peter,
>>
>> Thanks for your answer and the interesting thread you linked
>> (I should have catch it before sending my email ;-)).
>>
>> I agree that Platt scaling would be of help. As far as I understand
>> in t
2012/7/9 Philipp Singer :
> Am 09.07.2012 13:47, schrieb Peter Prettenhofer:
>> Hi,
>
> Hey!
>>
>> some quick thoughts:
>>
>> - if you use a multinomial Naive Bayes classifier (aka a language
>> model) you can fit a background model on the large dataset and use
>> that to smooth the model fitted on
Am 09.07.2012 13:47, schrieb Peter Prettenhofer:
> Hi,
Hey!
>
> some quick thoughts:
>
> - if you use a multinomial Naive Bayes classifier (aka a language
> model) you can fit a background model on the large dataset and use
> that to smooth the model fitted on the smaller dataset.
That's a nice i
Am 09.07.2012 13:59, schrieb Vlad Niculae:
> Another (hackish) idea to try would be to keep the labels of the extra
> data bit give it a sample_weight low enough not to override your good
> training data.
That's actually a great and simple idea. Would I do that similar to that
example:
http://sc
Another (hackish) idea to try would be to keep the labels of the extra
data bit give it a sample_weight low enough not to override your good
training data.
On 09.07.2012, at 12:43, Philipp Singer wrote:
> Hey!
>
> I am currently doing text classification. I have the following setup:
>
> 78 class
2012/7/9 Emanuele Olivetti :
> Hi Peter,
>
> Thanks for your answer and the interesting thread you linked
> (I should have catch it before sending my email ;-)).
>
> I agree that Platt scaling would be of help. As far as I understand
> in the multiclass case [0] it should be like:
>
> s = clf.decis
Hi,
some quick thoughts:
- if you use a multinomial Naive Bayes classifier (aka a language
model) you can fit a background model on the large dataset and use
that to smooth the model fitted on the smaller dataset.
- you should look at the domain adaptation / multi-task learning
literature - this
Hi,
You can try setting this as a semi-supervised learning problem and using
label propagation:
http://scikit-learn.org/stable/modules/label_propagation.html#label-propagation
HTH,
G
--
Live Security Virtual Conference
Hi Peter,
Thanks for your answer and the interesting thread you linked
(I should have catch it before sending my email ;-)).
I agree that Platt scaling would be of help. As far as I understand
in the multiclass case [0] it should be like:
s = clf.decision_function(x)
p(s) = exp(A*s+B) / exp(A*s+
Hey!
I am currently doing text classification. I have the following setup:
78 classes
max 1500 train examples per class
overall around 90.000 train examples
same amount of test examples
I am pretty happy with the classification results (~52% f1 score) which
is fine for my task.
But now I have
Hi Emanuele,
2012/7/9 Emanuele Olivetti :
> Hi,
>
> I'm not expert in Stochastic Gradient Descent but I'm wondering why
> predict_proba() is not available in the multiclass case. Precisely
> I get this (sklearn v0.11):
>
> NotImplementedError: predict_(log_)proba only supported for binary
>
Hi,
I'm not expert in Stochastic Gradient Descent but I'm wondering why
predict_proba() is not available in the multiclass case. Precisely
I get this (sklearn v0.11):
NotImplementedError: predict_(log_)proba only supported for binary
classification
I see that the multiclass case is supp
17 matches
Mail list logo
