Concepts de base
합리적인 Viterbi 디코딩을 위한 상호 정보 및 선형 최소 평균 제곱 오차 유도
Stats
zk = √ρ xk + wk, k = 0, 1, · · ·
E[wkwTl] = Wkδkl
E[νkνTl] = (Sk + Wk)δkl
Citations
"The covariance matrix of the innovation (denoted νk) is defined by νk △= zk − ˆsk|k−1."
"The mutual information I[xk; zk] between xk and zk is given by I[xk; zk] = 1/2 Σ log |Rj|/|Wj|."