A System of Subroutines for Iteratively Reweighted Least Squares Computations

A description of a system of subroutines to compute solutions to the iteratively reweighted least squares problem is presented. The weights are determined from the data and' linear fit and are computed as functions of the scaled residuals. Iteratively reweighted least squares is a part of robust statlstms where "robustness" means relative insensltlwty to moderate departures from assumptions. The software for iteratively rewelghted least squares is cast as semiportable Fortran code whose performance is unaffected (m the sense that performance will not be degraded) by the computer or operating-system environment in which it is used. An /~ start and an ~2 start are provided. Eight weight functions, a numerical rank determination, a convergence criterion, and a stem-and-leaf display are included.