AN UPPER BOUND ON THE NUMBER OF LINEAR RELATIONS IDENTIFIED FROM NOISY DATA BY THE FRISCH SCHEME

摘要

The Frisch scheme is one of many approaches to the identification of linear relations from noisy data. Solution of the associated matrix optimization problem leads to the identification of a maximal number of such relations. To date no general solution procedure is known. Here an upper bound on the maximal number of relations that can be identified is presented. The upper bound is derived using the theory of Schur complements in conjunction with a result for the case where only one relation can be identified. The possibility of developing a completely general solution procedure using these ideas is also briefly discussed. [References: 10]