Key Formulas of Chapter 9 - PowerPoint PPT Presentation

1 / 10
About This Presentation
Title:

Key Formulas of Chapter 9

Description:

yt(m) = E(et m f1 et m-1 . . . f1m-1 et 1 f1m yt ) = f1m yt ... Seasonal Nonstationary Prediction Interval: Yt = Yt-12 et ... – PowerPoint PPT presentation

Number of Views:46
Avg rating:3.0/5.0
Slides: 11
Provided by: foreca
Category:

less

Transcript and Presenter's Notes

Title: Key Formulas of Chapter 9


1
Key Formulas of Chapter 9
  • AR (1)
  • yt et f1 et-1 f12 et-2 . .. f1n et-p
    . . . (9-1)
  • Unconditional Forecast
  • E(Yt ) E(yt ) m m

2
  • Conditional Forecast
  • E(yt1 yt, yt-1, . . . , y2, y1) (9-3)
  • yt(m) E(etm f1 etm-1 . . . f1m-1 et1
    f1m yt ) f1m yt
  • Since f1 lt 1, f1m yt , ? 0 as m ? ?
    (9-5)
  • Then, Yt(m) yt(m) m 0 m m as m ?
    ?
  • Conditional Forecast ?
  • Unconditional Forecast as m ? ?

3
  • Forecast Mean Squared Error (FMSE) and Standard
    Error (FSE)
  • S Yt1m - Yt1 - (m)2
  • FMSE(m)
  • j
  • S Yt1m - Yt1 - (m)2
  • FSE(m)
  • j
  • EMSE (m) Expected FMSE(m)
  • EFSE(m) EMSE (m)

4
  • General ARIMA Models psi weights (y1)
  • yt et y1 et-1 y2 et-2 . .. yk et-k
    . . . (9-8)
  • (e.g., AR(1))
  • yt et f1 et-1 f12 et-2 . .. f1n et-p
    . . . (9-9)
  • General EMSE
  • EMSE(1) E(et12) se2 (9-15)
  • EMSE(2) se2 y12se2 E(et12) (9-19)
  • (1 y12 )se2
  • EMSE(3) (1 y12 y22)se2 (9-21)
  • EMSE(m) (1 y12 y22 . . . ym-12)se2
    (9-22)

5
  • P.I.
  • Yt (m) - tEFSE(m) ? Ytm ? Yt (m) tEFSE(m)
    (9-23)
  • AR(1)
  • EMSE(1) se2
  • EMSE(2) (1 f12 )se2
  • . . .
  • EMSE(m) (1 f12 f14 . . .
    f12m-2)se2 se2
  • EMSE(m) (9-25)
  • lim m ? ? (1 f12 )

6
  • ARIMA(0,1,0)
  • Yt Yt-1 et
  • Yt (1 B B2 B3 . . . Bn) et
  • y1 y2 . . . yk 1
  • Conditional Forecasts
  • Yt(m) E(Yt et1 . . . f1n etm )
    Yt (9-27)
  • EMSE(1) E(et12) se2
  • EMSE(2) E(et2 et1) 2 2se2
  • . . .
  • EMSE(m) E(etm etm-1 et1) 2
  • mse2 (9-28)

7
  • ARIMA (0,0,1), only one nonzero y weight, a
    one-period memory
  • yt et - q1 et-1 (1 q1B) et (1 y1B) et
  • where y1 -q1
  • thus, y1 y2 . . . yk 0 for all k gt 1

8
  • Conditional Forecasts
  • yt(1) E(et1 - q1et ) - q1et yt (1) m -
    q1et
  • yt(2) E(et2 - q1et1 ) 0 yt (2) m
  • . . .
  • yt(m) E(etm - q1etm-1 ) 0 yt (m) m
  • EMSE(1) E(et12) se2
  • EMSE(2) E(et2 - q1et1) 2 (1 q12) se2
  • . . .
  • EMSE(m) E(et2 - q1et m-1) 2 (1
    q12)se2 (9-29)
  • EMSE(m) (1 q12)se2
  • m ? ?

9
  • Seasonal Nonstationary Prediction Interval
  • Yt Yt-12 et
  • Yt (1 B12 B24 B36 . . . Bi12) et
    (9-30)
  • n
  • for i 1 to INT
  • 12
  • Yt (1 y12 B12 y24 B24 y36 B36 . . .
  • yi12 BI12) et (9-31)
  • y12 y24 . . . yi12 1,
  • all other yr 0 for r ? i12

10
  • Conditional Forecasts ARIMA(0,1,0)12
  • Yt(1) E(Yt-11 et1 ) Yt-11
  • Yt(12) E(Yt et12 ) Yt
  • Yt(13) E(Yt1 et13 ) E(Yt1) Yt(1)
    Yt-11
  • EMSE(1) se2
  • EMSE(2) (1 0) se2 se2
  • . . .
  • EMSE(25) (1 0 . . . 0 1 0 . . .
    0 1)se2
  • 3se2
  • m - 1
  • EMSE(m) INT 1 se2 (9-32)
  • 12
Write a Comment
User Comments (0)
About PowerShow.com