“…where x max is the column vector with unit norm that achieves the maximum matrix norm. Similarly, using the definition of the matrix norm again: A more detailed proof with all intermediate steps is provided in [37]. The value of sin(θ k,k (A, B)), defined earlier as T k,k (A, B) , is the first singular value of T k,k (A, B) and also the square root of the largest eigenvalue of T ′ k,k (A, B)T k,k (A, B).…”