I think SVM does not directly model "Probability" and optimize some
quantity regarding the probability like what logistic regression do. How to
get the probability highly depends on your project(you want high recall or
high precision or some balance over them). In upstream, libsvm does not
seem to
Hello, I am a new user of scikit-learn and I primarily use 2-class SVM and
1-class SVM. Now, I meet with a problem and hope someone can give me a reply.
I want the 1-class SVM probability output just like the function 2-class SVM
has provided. I know that scikit-learn implementation is based
