Estimation of Model Parameters
From Jane Harvill
Methods for estimating the coefficients in the simple linear regression model are investigated. The least squares method is a purely mathematical approach, requiring no statistical assumptions other than the function relating the response variable to the predictor variable is a linear function of the predictor variable. The method that yields best linear unbiased estimators requires statistical assumptions, but only on the first two moments of the response variable, or equivalently on the errors.