In this section, we derive and discuss the biance-variance decomposition of the generalization error of an inducer.