On the alternating projections theorem and bivariate stationary stochastic processes
Habib Salehi · Transactions of the American Mathematical Society · 1967
In this paper we shall first use the theorem of von Neumann on alternating projections to obtain an algorithm for finding the projection of an element x in a Hilbert space H \mathcal {H} onto the subspace spanned by H \mathcal {H} -valued orthogonally scattered measures ξ 1 {\xi _1} and ξ 2 {\xi _2} . We then specialize this algorithm to the case that ξ 1 {\xi _1} and ξ 2 {\xi _2} are the canonical measures of the components of a bivariate stationary stochastic process (SP), and thereby get an algorithm for finding the best linear predictor in the time domain.