Some situations in which maximum likelihood estimators do not have the standard properties of maximum likelihood...

 

Transcribed Image Text:

Some situations in which maximum likelihood estimators do not have the standard properties of maximum likelihood estimators are (i) One or more components of the true parameter are on the boundary (ii) The sample space depends on the parameter (iii) The model contains too many parameters. (iv) The specified model does not contain the true model. For 2 of these cases, create examples which illustrate the non-standard behaviour of the mle. Present your examples as questions set for fellow students and provide model solutions for both questions. Higher marks will be awarded for non-standard examples (e.g. not using the independent uniform[0, 0] model in (ii)), for simple examples which nonetheless demon- strate the point clearly, and for not simply using examples from the internet without any modification. You may search for examples in books or on the internet; if you do do this, please reference your source and indicate clearly any modifications you have made. Some situations in which maximum likelihood estimators do not have the standard properties of maximum likelihood estimators are (i) One or more components of the true parameter are on the boundary (ii) The sample space depends on the parameter (iii) The model contains too many parameters. (iv) The specified model does not contain the true model. For 2 of these cases, create examples which illustrate the non-standard behaviour of the mle. Present your examples as questions set for fellow students and provide model solutions for both questions. Higher marks will be awarded for non-standard examples (e.g. not using the independent uniform[0, 0] model in (ii)), for simple examples which nonetheless demon- strate the point clearly, and for not simply using examples from the internet without any modification. You may search for examples in books or on the internet; if you do do this, please reference your source and indicate clearly any modifications you have made.

Answer rating: 100% (QA)

a One or more components of the true parameter are on the boundary b The sample space depends on th... View the full answer