Thanks! It seems very counter-intuitive. If I'm calling fit(), it should
return the
best model after grid search (especially if I'm giving refit=True as a
parameter).
Unfortunately, printing the return value doesn't show the best estimator;
even
the get_params() call doesn't reveal it. One has to access the "hidden"
_best_estimator
parameter.
On Thu, Jan 22, 2015 at 5:33 PM, Joel Nothman <joel.noth...@gmail.com>
wrote:
> That's not the learnt estimator. You're looking at the initial input (i.e.
> the parameters that are or are not changed during the search). The learnt
> estimators are cloned from that one, and the best is stored at
> clf.best_estimator_ (if refit=True).
>
> Cheers, Joel
>
> On 23 January 2015 at 12:20, Aardvark Zebra <exma...@gmail.com> wrote:
>
>> I just started with s-l, and was playing around with it in iPython using
>> the Iris set.
>>
>> I created an SVM classifier thusly:
>>
>> clf = grid_search.GridSearchCV(svm.SVC(), param_grid={'kernel':('linear',
>> 'rbf'), 'C':arange(10,20)})
>>
>> (Basically, I want to grid-search for different parameters of "C", and 2
>> kernel functions).
>>
>> However, when I train it (fit) like this:
>> clf.fit(iris.data, iris.target)
>>
>> I get the following back:
>>
>> GridSearchCV(cv=None,
>> estimator=SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,
>> degree=3, gamma=0.0, kernel='rbf', max_iter=-1, probability=False,
>> random_state=None,
>> shrinking=True, tol=0.001, verbose=False),
>> fit_params={}, iid=True, loss_func=None, n_jobs=1,
>> param_grid={'kernel': ('linear', 'rbf'), 'C': array([10, 11, 12, 13, 14,
>> 15, 16, 17, 18, 19])},
>> pre_dispatch='2*n_jobs', refit=True, score_func=None, scoring=None,
>> verbose=1)
>>
>>
>> The learned estimator has "C" = 1.0 ; when the grid-search was for the range
>> 10 .. 19 (it's just an example...).
>>
>> Shouldn't the value of C be in this range?
>>
>>
>> Thanks.
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
>> GigeNET is offering a free month of service with a new server in Ashburn.
>> Choose from 2 high performing configs, both with 100TB of bandwidth.
>> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
>> http://p.sf.net/sfu/gigenet
>> _______________________________________________
>> Scikit-learn-general mailing list
>> Scikit-learn-general@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>
>
> ------------------------------------------------------------------------------
> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
> GigeNET is offering a free month of service with a new server in Ashburn.
> Choose from 2 high performing configs, both with 100TB of bandwidth.
> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
> http://p.sf.net/sfu/gigenet
> _______________________________________________
> Scikit-learn-general mailing list
> Scikit-learn-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
------------------------------------------------------------------------------
New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
GigeNET is offering a free month of service with a new server in Ashburn.
Choose from 2 high performing configs, both with 100TB of bandwidth.
Higher redundancy.Lower latency.Increased capacity.Completely compliant.
http://p.sf.net/sfu/gigenet
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
