Loss functions and estimation
In the previous chapter we reviewed some results of linear (least squares) models without making the loss function explicit. In this chapter we remedy this and extend the discussion to various other (sometimes referred to as "robust") approaches. That the loss function determines the properties of estimators is common to classical and Bayesian statistics (whether made explicit or not). We’ll review a few loss functions and the associated expected loss minimizing estimators. Then we briefly review maximum likelihood estimation (MLE) and nonlinear regression.
KeywordsMaximum Likelihood Estimation Posterior Distribution Loss Function Information Matrix Discrete Choice Model
Unable to display preview. Download preview PDF.