For predicting Y = (Yl, Yq)' from the value of x = (Xl, ... , Xp-d' we

Question:

For predicting Y = (Yl,"" Yq)' from the value of x = (Xl, ... , Xp-d' we say that a predictor f(x) is best if the scalar E{[y - f(x)]'[y - f(x)]} is minimized. Show that with simple modifications, Theorems 6.3.1 and 6.3.2 hold for the extended problem, as does Proposition 6.3.3.

Fantastic news! We've Found the answer you've been seeking!