{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

07_ts_chap5

# 07_ts_chap5 - Outline MSE prediction Box-Jenkins procedure...

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

Outline MSE prediction Box-Jenkins procedure Forecasting and Box-Jenkins Methodology Haipeng Xing, AMS316, Stony Brook University Forecasting and Box-Jenkins Methodology Outline MSE prediction Box-Jenkins procedure Outline 1 Mean square error prediction 2 Box-Jenkins procedure Haipeng Xing, AMS316, Stony Brook University Forecasting and Box-Jenkins Methodology

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

View Full Document
Outline MSE prediction Box-Jenkins procedure Mean square error prediction Suppose that the observations X 1 , . . . , X n follow some time series model. Denote F n = ( X 1 , . . . , X n ) . We want to find an estimate f ( X 1 , . . . , X n ) for X n + k ( k 1) , i.e., a k -step ahead forecast at forecast origin n . A common used forecast criterion is to consider minimizing the mean square error (MSE) of your forecast, i.e., min f E X n + k - f ( X 1 , . . . , X n ) 2 |F n = Var ( X n + k |F n ) + min f E E ( X n + k |F n ) - f ( X 1 , . . . , X n ) 2 |F n . The above MSE is minimized by choosing f = E ( X n + k |F n ) . Therefore, we refer to E ( X n + k |F n ) as the k -step ahead forecast .
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}