If the likelihood function (the prob. model) is correctly specified
Good: for "well-behaved" distributions, established large sample theory, asymptotically unbiased, asymptotic variance, optimal among asymp. unbiased estimates; likelihood ratio methods (tests, interval estimates).
Bad: likelihood surface may have many local maxima ; nuisance parameters
Immediately related elementsHow this works
-
Controlling Infectious Diseases Â»Controlling Infectious Diseases
▲
Research Questions Â»Research Questions
▲
Open mathematical/computational problems? Â»Open mathematical/computational problems?
▲
Statistical Methods for model selection and parameter estimation Â»Statistical Methods for model selection and parameter estimation
▲
Parametric approaches Â»Parametric approaches
▲
Maximum likelihood estimation Â»Maximum likelihood estimation
■
If the likelihood function (the prob. model) is correctly specified
◄
Sensitive to model misspecification Â»Sensitive to model misspecification
+Kommentare (0)
+Verweise (0)
+About