FORECAST Statement

FORECAST options ;

The FORECAST statement specifies the operations to be performed on the response series as well as the autoregressive and moving average polynomials in the model. The presence or absence of a constant in the model is also specified here.

The following options are used in the FORECAST statement.

SYMBOL=variable

VAR=variable

specifies a symbolic name for the dependent series. This symbol specification is optional. If the SYMBOL= option is not specified, Y is used as a default symbol.

DIF=order

DIF=( order1, order2, ...)

specifies the differencing orders for the dependent series. For example, DIF= (1 12) specifies that the series be differenced using the operator . The differencing orders can be positive integers or they can be "s", which indicates a placeholder that will be substituted later with an appropriate value. The use of placeholders is explained further in Example 3.3.

P=order

P=(lag, ..., lag ) ...(lag, ..., lag )

P=(lag, ..., lag )<> ...(lag, ..., lag )<>

specifies the autoregressive part of the model. By default, no autoregressive parameters are fit.

P=(l , l , ..., l ) defines a model with autoregressive parameters at the specified lags. P= order is equivalent to P=(1, 2, ..., order ).

A concatenation of parenthesized lists specifies a factored model. For example, P=(1,2,5)(6,12) specifies the autoregressive model

Optionally, you can specify multipliers after the parenthesized lists. For example, P=(1)(1)12 is equivalent to P=(1)(12), and P=(1,2)4(1)12(1,2)24 is equivalent to P=(4,8)(12)(24,48). These multipliers can either be positive integers or they can be "s", which indicates a placeholder that will be substituted later with an appropriate value. The use of placeholders in the multiplier specification is explained in Example 3.3.

Q=order

Q=(lag, ..., lag ) ...(lag, ..., lag )

Q=(lag, ..., lag )<> ...(lag, ..., lag )<>

specifies the moving average part of the model. By default, no moving average parameters are fit.

The manner of specification of the moving average part is identical to the specification of the autoregressive part described in the P= option.

AR=value ...

lists starting values for the autoregressive parameters.

MA=value ...

lists starting values for the moving average parameters.

NOCONSTANT

NOINT

suppresses the fitting of a constant (or intercept) parameter in the model. (That is, the parameter is omitted.)

MU=value

specifies the MU parameter.

NOISEVAR=value

specifies the noise variance. This is useful only if you want to specify an externally published model that is fully specified.

TRANSFORM=option

specifies the transformation to be applied to the time series. The following transformations are provided:

NONE

no transformation applied

LOG

logarithmic transformation

SQRT

square-root transformation

LOGISTIC

logistic transformation

BOXCOX(n )

Box-Cox transformation with parameter number where number is between –5 and 5

When the TRANSFORM= option is specified, the intended time series must be strictly positive.

TRANSOPT=option

specifies that mean or median forecasts be estimated. The following options are provided:

MEAN

Mean forecasts are estimated. This is the default.

MEDIAN

Median forecasts are estimated.

If no transformation is applied to the actual series with the TRANSFORM= option, the mean and median time series forecast values are identical.