2015-05-22 9:01 GMT+02:00 zhenjiang zech xu :
> I tested the following code and its outputs show predict_proba and predict
> give very different result, even for the samples with high probability (0.7)
> to be label 1 are predicted as label 1. I am very surprised. Is this problem
> specific to the
As I know, SVM predicts class label directly without probability
information.
The probability of each class is just an estimation.
(Check section 8 of libsvm paper:
http://www.csie.ntu.edu.tw/~cjlin/papers/libsvm.pdf)
On Fri, May 22, 2015 at 3:03 AM zhenjiang zech xu
wrote:
> Hi all,
>
> I test
