English: This figure illustrates the relationship between overfitting and the generalization error I[f_n] - I_S[f_n]. In the left column, a set of training points is shown in black, while another sample from the same iid distribution is shown by the black points in the right column. The samples are drawn from a noisy version of the relationship y = x^2. In the top row, a quadratic polynomial is fit to the training sample, while in the bottom row, a fifth-order polynomial is fit. Overfitting takes place when the fifth-order polynomial is fit. While the empirical error on the training set is low, the expected error is high, as it is if another sample is used.