X is a covariance matrix of a gaussian vector R with a zero mean. Let Q=sign(R), that is Q_i=1 when R_i >=0 and Q_i=-1 when R_i < 0. Prove that E(Q_iQ_j)=(2/pi)(arcsin(X_i,j/(sqrt(X_i,i)sqrt(X_j,j))))?

More Maksim Sergeevich Barketau's questions See All
Similar questions and discussions