JamesTuna

The line of code is in signal_matrix.py.

I mean sqrt[sum(diff ^2)]/(2n).

I don't think they are the same. Isn't np.std(M) for computing standard deviation over all entries of matrix M? For example, if M1-M2 is a matrix with entries being all...

Oh I see, thanks for the explanation. Another thing confused me a lot is that why the variance for entries in M1 and M2 is double the variance of that...

Sure, I know what you just said. I think you misunderstand my question. I know variance of (M1-M2) is twice that of M1(also M2) but I don't know why variance...

To make my question clear: why variance of E1 or E2 is 2 sigma square? Where sigma square is variance for matrix computed using all samples.

Thanks, I will think through the variance problem later. Regarding matrix symmetry problem in class MonteCarloEstimator, I have several questions. 1. Why you generate X by (U_gen * D_gen).dot(V_gen.T)? V_gen...

For variance of error problem, I have trouble thinking through a simply constructed problem. Suppose X1 X2 ... iid random variables. Consider two cases: 1). Y(n) = (X1+X2+...+Xn) / n...