An investigation of certain classes of data-smoothing procedures for estimating a set of parameters on the basis of observed data which depend on these parameters and are also corrupted by noise. The procedures are in part motivated by possible appl...