A note on the statistical estimation of a hyperplane

Several authors have discussed the problem of statistically estimating the pulse transfer function of a linear system, or what is equivalent, estimating the vector normal to a hyperplane. Certain authors have raised objections to the least-squares estimator commonly proposed for this problem. This correspondence points out that these objections pertain not so much to the least-squares method per se as they do to an unfortunate choice of normalization.