Statistics and Probability

Show that the expected value of the empirical risk of a given classifier - namely, the number of instances in the training data set that the classifier assigns incorrect label to, equals the generalization error of this classifier - the probability that the classifier makes an error

Please solve fast.