# X12ARIMA - Defining an X12-ARIMA Model

Returns a unique string to designate the specified X12-ARIMA model.

## Syntax

X12ARIMA(X, Date, Period, Transform, Cal_Reg, Outliers, ARIMA, Forecast, SA_Filter, X11Options, X11Mode)

X is the univariate time series data (one dimensional array of cells (e.g. rows or columns)).

Date is a serial number that represents the data start date.

Period is the sampling rate per year ( 12 = monthly, 4 = Quarterly)

Transform is an option for transforming the data prior to analysis (1=Log, 2=Auto (auto), 3=None ). If missing, auto is assumed.

Type Desc
1 Log
2 Auto (auto)
3 None

Cal_Reg is a set of options for calendar adjustments using regression ({TD=0/1, Easter=0/1, Constant=0/1}). See the reference manual for more details and examples.

Type Examples
{0,0,0} TD=0, Easter=0, Constant=0 (default)
{1,0,0} TD=1, Easter=0, Constant=0
{1,1,1} TD=1, Easter=1, Constant=1
{1,1,1} TD=1, Easter=1, Constant=1

Outliers is a set of outlier types to test and adjust for {AO=0/1, LS(Run)=0/1..N, TC=0/1, SO=0/1, Hard-code=0/1}. If missing, no check for outliers is carried out.

Type Examples
{0,0,0,0} AO=0, LS=0, TC=0 (default)
{1,1,1,1} AO=1, LS=1, TC=1
{1,4,0,0} AO=1, LS=4, TC=0

ARIMA is the set of orders for the ARIMA Model ({p,d,q,P,D,Q}). If any of the orders is negative or if all values are zeros, auto-select is turned on.

Type Examples
{0,0,0,0,0,0} p=0, d=0, q=0, P=0, D=0, Q=0 (default Auto-select)
{-1,1,1,0,1,1} p=0, d=0, q=0, P=0, D=0, Q=0 (Auto-select)
{0,1,1,0,1,1} p=0, d=1, q=1, P=0, D=1, Q=1

Forecast is number of years to conduct the forecast for. If missing, forecast = 1 year.

SA_Filter is a flag for the seasonal adjustment filter to use (1=X11, 2=TD and Holidays, 3=None (default)).

Type Desc
1 X11
3 None (default)

X11Options is an option setting for the X11 filter (1=X11 default, 2=3x1, 3=3x3, 4=3x5, 5=3x9, 6=3x15 , 7=Stable ).

Type Desc
1 X11 default
2 3x1
3 3x3
4 3x5
5 3x9
6 3x15
7 X11 Stable

Type Desc
1 Multiplicative default

## Remarks

1. The underlying X12-ARIMA model is described here.