Minimizing mean-squared error for random sequences

rsingh628

New member
Joined
May 31, 2021
Messages
42
Hi all, I’m stuck on where to begin. For the question below, my approach is simply to substitute the Sk(hat)S_k (hat) into the MSE formula, get an expression in terms of c0 c_0 and take the derivative and setting it to 0 to find the min c0 c_0 . I guess that is where I’m stuck in that I have the MSE in terms of expectations. If I substitute Sk(hat)S_k (hat) into the formula and distribute terms, I get MSE=(c0)2E(Xk2)2c0E(XkSk)+E(Sk2)MSE = (c_0)^2E(X_k^2) - 2c_0E(X_kS_k) + E(S_k^2) so I’m struggling to find of those expectations. The Sk S_k are iid Gaussian with mean 1 and variance 1, but when I try to calculate, say, E(Xk) E(X_k) , I get 0 since all Sk S_k are independent. That doesn’t help much, so I guess my other issue is finding E(Xk2)E(X_k^2), E(XkSk)E(X_kS_k), E(Sk2)E(S_k^2) etc.

Any help would be much appreciated, thank you.


1667744664957.png
 
Last edited:
Top