ABSTRACT
For many regression models and time series models that assume nor-
mality of the noise distribution, least squares estimates may often coin-
cide with or provide good approximations to the maximum likelihood
estimates of the unknown parameters. This chapter explains the House-
holder transformation as a convenient method to obtain least squares es-
timates of regression models (Golub (1965), Sakamoto et al. (1986)).
With this method, we can obtain precise estimates of the coefficients of
the model and perform order selection or variable selection based on the
information criterion AIC quite efficiently.
