Hello all,
I have three general questions regarding generating a confidence percentage on
a classification prediction:
1. can calibration / brier score loss, be used for a multi-class
classification? The scikit-learn page
(http://scikit-learn.org/stable/auto_examples/calibration/plot_calibration.html)
doesn't restrict it to a binary case, but the wiki
(https://en.wikipedia.org/wiki/Brier_score) seem to suggest binary.
2. in `y_prob = np.array([0.1, 0.9, 0.8, 0.3])`, how was the probability
for each categorical target obtained -
http://scikit-learn.org/stable/modules/generated/sklearn.metrics.brier_score_loss.html#sklearn.metrics.brier_score_loss?
It only an example, but how would it be generated? I'm curious because I'm
not sure how to use the `brier_score_loss` after the `predict_proba`
implementation
-https://github.com/jeff1evesque/machine-learning/issues/1924#issuecomment-159491052
3. does svc have a built in method, to check if a dataset is sparse
enough? If not, is there a way I can script a check, indicating the
corresponding dataset is not large enough to determine a prediction, or is this
the brier score loss?
Ultimately, I’m trying to find ways to implement the `predict_proba` method,
with an accurate corresponding confidence percentage
-https://github.com/jeff1evesque/machine-learning/issues/1924#issuecomment-160847135
Thank you,
Jeffrey Levesque
https://github.com/jeff1evesque/
(603) 969-5363
Sent from my iPhone
------------------------------------------------------------------------------
Go from Idea to Many App Stores Faster with Intel(R) XDK
Give your users amazing mobile app experiences with Intel(R) XDK.
Use one codebase in this all-in-one HTML5 development environment.
Design, debug & build mobile apps & 2D/3D high-impact games for multiple OSs.
http://pubads.g.doubleclick.net/gampad/clk?id=254741911&iu=/4140
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
