Lecture 20 - Measures of the Estimators

Lecture 20 - Measures of the Estimators - 1 MEASURES OF THE...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 1 MEASURES OF THE ESTIMATORS Lecture 20 ORIE3500/5500 Summer2011 Li 1 Measures of the Estimators A point estimator of a parameter θ of the distribution of the observations X 1 ,...,X n is any suitable statistic ˆ θ = ˆ θ ( X 1 ,...,X n ) and a value obtained by evaluating the function for the observed values of the random variables is a point estimate . We have already discussed that an estimator ˆ θ is unbiased for a parameter θ if E £ ˆ θ ( X 1 ,...,X n ) / = θ, otherwise ˆ θ is a biased estimator of θ . For a biased estimator ˆ θ of the pa- rameter θ , the bias of ˆ θ is defined as bias ( ˆ θ ) = E £ ˆ θ ( X 1 ,...,X n ) /- θ. So bias is a measure of how good the estimator is. Another such measure is consistency. An estimator ˆ θ is consistent for θ if, for any ² > 0, P ( | ˆ θ ( X 1 ,...,X n )- θ | > ² ) → ,n → ∞ . We also say that ˆ θ converges to θ in probability. This means that for large n we can expect that ˆ θ ( X 1 ,...,X n ) is close to θ for large values of n . The mean squared error is another measure for the precision of an esti- mator. The name is self explanatory MSE ( ˆ θ ) = E £ ( ˆ θ ( X 1 ,...,X n )- θ ) 2 / Facts: For any estimator ˆ θ of θ , MSE( ˆ θ ) = Bias( ˆ θ ) 2 + Var( ˆ θ ) Proof. MSE( ˆ θ ) = E ( ˆ θ- θ ) 2 = E [ { ˆ θ- E ( ˆ θ ) } + { E ( ˆ θ )- θ } ] 2 = E { ˆ θ- E ( ˆ θ ) } 2 + { E ( ˆ θ )- θ } 2 + 2 { E ( ˆ θ )- θ } E { ˆ θ- E ( ˆ θ ) } = Var( ˆ θ ) + Bias( ˆ θ ) 2 1 2 MAXIMUM LIKELIHOOD ESTIMATES where the last step follows since 2 { E ( ˆ θ )- θ } E { ˆ θ- E ( ˆ θ ) } = 2 { E ( ˆ θ )- θ }{ E ( ˆ θ )- E ( ˆ θ ) } = 0 Intuitively, comparing the MSE is the most natural way to compare two estimators ˆ θ 1 and ˆ θ 2 , i.e....
View Full Document

This note was uploaded on 12/09/2011 for the course IMSE 0301 taught by Professor Song during the Spring '11 term at HKU.

Page1 / 6

Lecture 20 - Measures of the Estimators - 1 MEASURES OF THE...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online