The interpretation of maximum-likelihood estimation

From MaRDI portal





If an estimator of an unknown parameter \(\theta\) is to be based on an observed random vector X, the authors suggest that maximum likelihood estimation should be interpreted as a procedure for generating quantities u(X;\(\theta)\) which have the following properties: (1) u(X;\(\theta)\) is at least approximately pivotal, that is, its distribution is at least approximately independent of \(\theta\) ; (2) u(X;\(\theta)\) is efficient in the sense that it adequately summarizes the information in X by reproducing the original likelihood function of \(\theta\) based on X; (3) u(X;\(\theta)\) should be approximately linear in \(\theta\).











This page was built for publication: The interpretation of maximum-likelihood estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3330299)