SPECFORE Procedure

@SPECFORE computes forecasts for a single series using spectral techniques.

@SPECFORE( options ) series start end forecasts

Parameters

series	(input) series to be forecast
start, end	range of entries to forecast
forecasts	(output) series for computed forecasts

Options

The options choose how the data are transformed prior to the analysis; the goal is to produce a stationary series. It's the underlying stationary series that gets forecast, and forecasts for the original data are created by backtransforming.

DIFFS=Number of regular differences [0]

SDIFFS=Number of seasonal differences [0]

[CONST]/NOCONST

TRANS=[NONE]/LOG/ROOT

Transformation to apply to data

Example

open data haversample.rat

calendar(m) 1960

data(format=rats) 1960:1 2007:4 fcm5

* Compute a forecast for the last 16 months of actual data

@specfore(diffs=1) fcm5 2006:1 2007:4 insample

* Graph the forecasts and actuals:

graph(key=below,header="Interest Rate Forecast") 2

# fcm5 2006:1 2007:4

# insample 2006:1 2007:4

* Compute an 20-month out-of-sample forecast and graph the results:

@specfore(diffs=1) fcm5 2007:5 2008:12 outofsample

graph(key=below,header="Interest Rate Forecast") 2

# fcm5 2006:1 2007:4

# outofsample