post10

# post10 - Quality of Estimators Example f(x = 1/2 Mean...

This preview shows pages 1–2. Sign up to view the full content.

Quality of Estimators Example f ( x )= 1 / 2 - 100 <x<θ - 99 , 1 / 2 +99 + 100 , 0 otherwise Let u ( x x . Then, it is an unbiased estimator of θ . But will it ever be close to θ ? Unbiasedness is not enough. 1 Mean Squared Error (MSE) The mean square error(MSE) of an estimator ˆ θ = u ( X 1 , ··· ,X n )fo r θ is: MSE( ˆ θ E (( ˆ θ - θ ) 2 ) =var( ˆ θ )+(Bias( ˆ θ )) 2 A small MSE is a good property for an estima- tor to have. 2 Example : X 1 , n : iid N ( μ, σ 2 ) Consider two estimator of σ 2 : S 2 n = 1 n X ( X i - ¯ X ) 2 S * 2 n = 1 n - 1 X ( X i - ¯ X ) 2 . How do these compare in terms of MSE? 3 MSE and Risk MSE is a speciﬁc example of a more general concept called risk. Risk is deﬁned as the ex- pected loss: R ( θ, u ( X 1 , n )) = E (Loss( θ, u ( X 1 , n )) Loss function measures the distance between the estimate and the parameter. (e.g. Loss( θ, u ( u - θ ) 2 ) Can we ﬁnd the estimator with the smallest risk? With the squared error loss, MSE and risk are the same. Can we ﬁnd the estimator with the

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

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

### Page1 / 2

post10 - Quality of Estimators Example f(x = 1/2 Mean...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online