gt-ln04

# gt-ln04 - LN04: FORECASTING YASUTOMO MURASAWA Contents 1....

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

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: LN04: FORECASTING YASUTOMO MURASAWA Contents 1. Forecasting 1 1.1. Statistical Decision 1 1.2. Best Prediction 1 1.3. Linear Projection 2 1.4. OLS 3 2. Forecasts with Infinite Observations 3 2.1. MA Sequences 3 2.2. ARMA Sequences 4 3. Forecasts with Finite Observations 4 Appendix A. LDU Decomposition 4 1. Forecasting 1.1. Statistical Decision. Let ( y t , x t ) be a (1 + k )-variate random vector. Let y e t be a prediction of y t , g ( . ) be a predictor of y t . Definition 1. A loss function is L ( ., . ) that gives the loss of using y e t instead of y t . Definition 2. A quadratic loss function is L ( ., . ) s.th. ( y e , y ) R 2 , L ( y e , y ) ( y e- y ) 2 . Definition 3. The risk of g ( . ) is R ( g ( . )) := E( L ( g ( x t ) , y t )) . Definition 4. The mean squared error (MSE) of g ( . ) is MSE( g ( . )) := E ( ( y t- g ( x t )) 2 ) . Definition 5. The MSE is the risk under a quadratic loss. 1.2. Best Prediction. Let G be the set of all predictors of y t . Definition 6. The best predictor of y t is g ( . ) := arg min g ( . ) G R ( g ( . )) . Theorem 1. Suppose that (1) L ( ., . ) is quadratic, (2) ( y t , g ( x t )) is square-integrable. Then g ( x t ) = E( y t | x t ) . Date : October 28, 2009. 1 2 YASUTOMO MURASAWA Proof. By the LIE, g ( . ) G , E ( ( y t- g ( x t )) 2 ) = E ( [( y t- E( y t | x t )) + (E( y t | x t )- g ( x t ))] 2 ) = E ( ( y t- E( y t | x t )) 2 ) + 2 E(( y t- E( y t | x t ))(E( y t | x t )- g ( x t ))) + E ( (E( y t | x t )- g ( x t )) 2 ) E ( ( y t- E( y t | x t )) 2 ) + 2 E(E(( y t- E( y t | x t ))(E( y t | x t )- g ( x t )) | x t...
View Full Document

## This note was uploaded on 02/22/2012 for the course EE 441 taught by Professor Neely during the Spring '08 term at USC.

### Page1 / 6

gt-ln04 - LN04: FORECASTING YASUTOMO MURASAWA Contents 1....

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

View Full Document
Ask a homework question - tutors are online