On Homogeneous least-squares . . .

The term "homogeneous least-squares" refers tomodel of the form Ya 0, where Y is some data matrix, and a is an unknown parameter vector to be estimated. SuchproblGk are encountered, e.g., whenmodelHL auto-regressive (AR) processes. Naturals. in order toappl a lGGNqTkHH#Lq (LS)solqxUx to suchmodelk the parameter vector a has to be somehow constrained in order to avoid thetrivial solial a = 0.UsualNq theprobl= at hand lndq to a "natural constraint on a. However, itwil be shown that the use of somecommonl applnl constraints, such as a quadratic constraint, can lnq to inconsistent estimates of a