On Fri, Sep 30, 2011 at 7:25 PM, Gael Varoquaux
<[email protected]> wrote:

> I would try using the predict_proba method of the LogisticRegression
> which give the probability of each class, and threshold it at a different
> value than .5, i.e. bias the decision.

This seems very similar to learning... a bias term
(fit_intercept=True). Unfortunately, In my experience, carefully
learning the bias term can be very hard and some algorithms do quite
bad at learning it. I would need to check this assertion but it seems
to me that SVC does a better job at learning the bias term than
LinearSVC.

Mathieu

------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security
threats, fraudulent activity, and more. Splunk takes this data and makes
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2dcopy2
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to