A detailed discussion of the calculation of the "best straight line" by the method of least squares is given. The most general solution is found and the conditions under which certain previously derived special solutions are valid are clearly stated. The "best" slope is shown to be given by the solution of the "Least-Squares Cubic". An example is given to illustrate the method. It is shown that the best slope is not necessarily bounded by values found from the regressions of x on y and y on x.
No takes yet. Share an insight, caveat, or question.
Derek York (1966) studied this question.
Synapse has enriched one closely related paper. Consider it for comparative context: