List the advantages and disadvantages of maximum likelihood methods
Advantages.
- Simple to apply
- Lower variance than other methods (i.e. estimation method least affected by sampling error) and unbiased as the sample size increases.
- The method is statistically well understood
- Able to analyze statistical models with different characters on the same basis. Maximum likelihood provides a consistent approach to parameter estimation problems. This means that maximum likelihood estimates can be developed for a large variety of estimation situations.
- Once a maximum-likelihood estimator is derived, the general theory of maximum-likelihood estimation provides standard errors, statistical tests, and other results useful for statistical inference.
Disadvantages.
- Computationally intensive and so extremely slow (though this is becoming much less of an issue)
- Frequently requires strong assumptions about the structure of the data
- The estimates that are obtained using this method are often biased. That is, they contain a systematic error of estimation. This is true for small samples. The optimality properties may not apply for small samples.
- MLE is inapplicable for the analysis of non-regular populations (Non-regular distributions are models where a parameter value is constrained by a single observed value)
********************
No comments:
Post a Comment