Yes, that assumption is correct (at least for now), but cv_scores_ also
gives you the parameters for each candidate.
If your parameter grid is actually a grid (not a list of grids), you can
reshape the data so that each parameter name corresponds to an axis of an
array by using this patch:
https://github.com/scikit-learn/scikit-learn/pull/1842
If you only want to compare the means and standard deviations you could
export it to pandas:
list_of_dicts = []
for candidate in search.cv_scores_:
d = candidate.parameters.copy()
d['_mean'] = candidate.mean_validation_score
d['_std'] = np.std(candidate.cv_validation_scores)
list_of_dicts.append(d)
frame = pandas.DataFrame(d)
- Joel
On Thu, Jul 18, 2013 at 5:21 AM, Josh Wasserstein <[email protected]>wrote:
> I would like to do some in-depth analysis of the distribution of scores
> for each set of parameters in my parameter grid.
>
> Is it safe to assume that the order of parameters in this list:
>
> *list_of_parameters* =
> list(sklearn.grid_search.ParameterGrid(my_parameter_grid)
>
> is the same order in which scores are returned in:
>
> clf = sklearn.grid_search.GridSearchCV(SVC, my_parameter_grid, ...)
> clf.fit(X,y)
>
> *list_of_scores_for_each_parameter* = [x.cv_validation_scores for x in
> clf.cv_scores_]
>
> This is all with 0.14-git
> ?
>
> Josh
>
>
> ------------------------------------------------------------------------------
> See everything from the browser to the database with AppDynamics
> Get end-to-end visibility with application monitoring from AppDynamics
> Isolate bottlenecks and diagnose root cause in seconds.
> Start your free trial of AppDynamics Pro today!
> http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
