1. Show that the eigenvalues of the matrix Evy EyxE Exy Eyy necessarily lie between zero and one. 2. Consider the weighted sum-of-squares reduced rank regression objective function W (t) = E(x,y) lly - M - A(t) B() X 1/?. Show that the above criterion is invariant under nonsingular transformations of the form X' = \\U + AX and Y' = $ + AY for any invertible matrices A and A with appropriate dimensions