Last squares methods for locating function shifts and disturbance regions

Presents some generalizations of a least squares method previously introduced by R. S. Bogartz to fit 2 intercepting lines to a set of data points. Polynomials of arbitrary order are considered instead of straight lines. The concept of a disturbance region is introduced. Methods are given for locating the disturbance region described by 1 polynomial in a portion of the range of the independent variable while another polynomial relates the dependent variable to the independent variable in the rest of the range of the independent variable. A decision rule for deciding the order of the polynomials under consideration, based on a sequence of F tests, is given.