Documente Academic
Documente Profesional
Documente Cultură
Gloria Gonzlez-Rivera
University of California, Riverside
and
Jess Gonzalo U. Carlos III de Madrid
Spring 2002
Forecasting in Action
Forecasts are made to guide decisions in a variety of fields.
Once you have done your forecast, someone else can come with
another forecast of the same variable. How do you compare these
forecasts?
Forecast Evaluation: Different measures of the forecast
errors.
y t 0 1x t t
t is WN(0, 2 )
y t 0 1x t t
t is N(0, 2 )
our h-step ahead conditional forecast for y, given that the h-step
value of x is x *T h is
y T h , T | x *T h 0 1x *T h
Assuming normality, we use the conditional density forecast
N ( y T h, T | x *T h ), and from it we get conditional interval forecasts.
We make the procedure operational by replacing unknown
parameters with estimates.
y t x t t
y T h, T | x *T h x *T h
T h , T y T h y T h , T ( ) x *T h T h
Thus,
var( T h , T ) ( x *T h ) 2 var( ) 2
y T h , T 0 1x T h , T
We usually don`t have such a forecast for x and the regression model
at hand doesnt help us. Assuming this variable follows and ARIMA
representation, you will learn how to produce these forecasts in the
next set of slides: FORECASTING II
Evaluation of Forecasts
There are many ways of making forecasts, but all of them need the
following common ingredients in order for success:
(i) that there are regularities to capture
(ii) that such regularities are informative about the future
(iii) they are encapsulated in the selected forecasting method, and
(iv) non-regularities and excluded.
The main alternatives are (for some of them see Reading I):
1) Guessing
2) Extrapolation
3) Leading Indicators
4) Surveys
5) Time-Series Models
6) Econometric Models
RMSE
1
MAE
T
t h, t
t 1
1
T
e2
t h, t
t 1
t 1
| e t h, t |
Forecast combination
be two forecasts of yt with errors
( j)
( j)
e t y t f t for j 1,2, and such that
( j)
( j)2
E (e t ) 0, E (e t ) 2 , and
j
(1) (2)
E (e t e t ) 1 2
(1)
(2)
C t kf t (1 k )f t
(c)
(1)
( 2)
e t y t C t ke t (1 k )e t
1 2
c2,0
2 2 21 2
1
2
2 min( 2 , 2 ) , unless 1 or 2
c,0
1 2
2
1
Note that
. If either
equality holds, then the variance of the combined forecast is equal to
the smaller of the two error variances.
2
k o 0 if and only if
1