How to modify Scipy's SVM classification rules?

Question

I have been exploring Scikit-learn as a tool, and I am very interested in learning if I can modify how Scikit-learn classifies a data point, more specifically, its SVM funct