Given a set of k "training samples" (X₁,Y₁), where X₁ is an m by k matrix and Y₁ an n by k matrix, and a set (X₂,Y₂) of k testing samples of the same dimensions, write a procedure (or give a series of matlab steps) to find the weight matrices W₁ and W₂, so that the norm(F(X₁) - Y₁) is minimized, and compare this with the norm(F(X₂) - Y₂).

A matlab program is provided that implements a sample non-linear function to generate test data sets (X₁,Y₁) and (X₂,Y₂); typically, X₁ and X₂ can be random, with Y_i the output of the sample non-linear function on input X_i. For this exercise, you can assume m is no larger than 8, n is no larger than 4, and k no larger than 20, though you can use larger values of k, if desired, to get a better fit.