ACF Procedure

@ACF performs an autocorrelation analysis on a series. This is a simpler procedure than the related @BJIDENT and @REGCORRS.

@ACF( options ) series start end

Parameters

series	series to analyze
start, end	range of series to use. By default, the defined range of series.

Options

NUMBER=number of autocorrelations to compute[depends upon number of data points]

METHOD=YULE/[BURG]

This chooses the algorithm used for computing the autocorrelations. BURG is generally thought to be more accurate; YULE is simpler and more commonly used. If you're trying to reproduced published results, it's more likely they were done with YULE.

REPORT/[NOREPORT]

[GRAPH]/NOGRAPH

You can graph the autocorrelations and produce a table (with REPORT) or both. The default is to graph only.

[QSTATS]/NOQSTATS

Computes and displays Ljung-Box Q Statistics

DFC=degrees of freedom correction for Q statistics

TITLE="title of graph window and report window"

HEADER="header string for graph"

FOOTER="footer string for graph"

Example

* Delurgio, Forecasting Principles and Applications

* Example BIRTHMAR.DAT from pp 160-161

open data birthmar.dat

calendar(q) 1985

data(format=free,org=columns) 1985:1 1992:4 date birth marriages

@acf marriages

* The seasonal smoothing is done with ESMOOTH with the option

* SEASONAL=ADDITIVE. To get the model shown, you set the level smoothing

* parameter (called ALPHA) to 0, and the seasonal smoothing parameter

* (called DELTA) to the "alpha" value described (here .435).

esmooth(seasonal=additive,alpha=0,delta=0.435,initial=start,fitted=sesfore) marriages

@uforeerrors marriages sesfore 5 *

set residuals = marriages-sesfore

@acf(report) residuals 5 *

Sample Output

This is from the instruction in the example. The first is the graph; note the Q-statistic inserted into the bottom left. The second is the report window. This has the autocorrelations, partial autocorrelations and running Q statistics (that is, the LAG 1 row has the Q for lag 1, the LAG 7 row has the Q for the first 7 autocorrelations).