{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

slides_5_conddist

# ∙ if we define “best” to be minimum mean square

This preview shows pages 53–60. Sign up to view the full content.

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

View Full Document

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

View Full Document

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

View Full Document

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.

Unformatted text preview: ∙ If we define “best” to be minimum mean square error then E Y | X is the best we can do. 53 ( ce8 ) (Minimum MSE Predictor) Among all functions of X , the conditional expectation minimizes the mean squared error. To be precise, let x ≡ E Y | X x . We want to use the outcome on X to predict Y , so our prediction functions can be written as g x . But we need to consider the random variable g X because we want a criterion that considers all possible outcomes on X . (For example, if X contains high school GPA and standardized test scores, we need to allow that X takes on many different values in the pool of applicants.) 54 ∙ Define MSE g E Y − g X 2 , where the expected value is over the joint distribution of X , Y . For the conditional mean , MSE E Y − X 2 55 ∙ The formal claim is that, if E Y 2 , then for any function with E g X 2 , E Y − X 2 ≤ E Y − g X 2 . In fact, we can say even more. Write Y − g X 2 Y − X X − g X 2 Y − X 2 2 Y − X X − g X X − g X 2 56 Take the expected value conditional on X : E Y − g X 2 | X E Y − X 2 | X 2 X − g X E Y − X | X X − g X 2 Var Y | X X − g X 2 because E Y − X | X 0 and E Y − X 2 | X Var Y | X . Now take the expectation and use the LIE: E Y − g X 2 E Y − X 2 E X − g X 2 ≥ E Y − X 2 57 2 . 2 . Properties of Conditional Variances and Covariances ( cv1 ) For any function a x , Var a X | X 0. ( cv2 ) For functions a x and b x , Var a X Y b X | X a X 2 Var Y | X ( cv3 ) The unconditional variance of Y can be expressed as Var Y E Var Y | X Var E Y | X ≡ E X Var Y | X Var X E Y | X ≡ E X 2 X Var X X 58 Proof: Let Y E Y and X E Y | X . Then Y 2 E Y − Y 2 ....
View Full Document

{[ snackBarMessage ]}

### Page53 / 67

∙ If we define “best” to be minimum mean square error...

This preview shows document pages 53 - 60. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online