The Statistical Estimation of a Rectangular Hyperbola

Publisher Website
Google Scholar
Abstract
A simple method is described for fitting a function of the type (X - a) (Y - b) = c to a set of observations by minimizing a quantity that closely approximates to the sum of the squared standardized normal residuals. It is also shown that certain difficulties hitherto associated with the maximum likelihood solution to a relation between variables all of which are liable to experimental error are due to confusion between the root-mean-square standardized residual and the root-mean-square standardized normal residual.