I trained a model for classification in SVC which I want to visualize the trainng process so the verbose is set to True. And this is the result I get:
optimazation finished,#iter = 200
obj = -315.082675, rho = 3.537407
nSV = 232, nBSV = 199
The iter says that the iteration needed for the convergence. What does the obj and rho mean in this case?
scikit-learn's SVC implementation is based on libsvm. You are trying to solve the ν-Support Vector Classification problem which is defined here as the following quadratic program (the primal problem):
where the corresponding dual problem is
The ν-SVC of a soft margin variant of the optimal hyperplane uses the ν-parameterization. The standard regularization parameter C is replaced by a parameter ν ∈ [0, 1] which is the lower and upper bound on the number of examples that are support vectors and that lie on the wrong side of the hyperplane, respectively, as mentioned here.
The variable ρ is optimized as well. Note that for ξ = 0, the two classes are separated by the margin 2ρ/||w||. Here rho refers to the optimal value of ρ at convergence. And obj is the optimal value of the objective function (which is minimized) at convergence.