(10 points) Fitting with continuous and discontinuous piecewise-linear functions. Consider a fitting problem with n= 1, so X(1), ..., X(N) and y(1), ..., y(N) are numbers. We consider two types of closely related models. The first is a piecewiselinear model with knot points at -1 and 1, as described on page 256, and illustrated in figure 13.8. The second is a stratified model (see page 272), with three independent affine models, one for x <: one for x> 1. (In other words, we stratify on X taking low, middle, or high values.) Are these two models the same? 15 one more general than the other? How manyr parameters does each model have? Hint. See problem title. What can you say about the training set RMS error and test set RMS error that would be achieved using least squares with these two models