As the title say i have my svm solution with kernel in the form of a few vector of lagrangian variables (alpha,gamma,delta) specific for the svm i'm working with, i have a vector (generated by a regressor but that's not important) and i need to calculate the angle between them.
Obviously i don't have the explicit vector of weights because i would need to know the mapping phi(x) but i don't know it due to the kernel trick.
How should i proceed?
[link][1 comment]