Autoregressive Moving-Average Generator

random seed

AR terms

0

1

2

3

4

5

weak stationarity condition

T

200

MA terms

0

1

2

3

4

5

approximate cut-off point bounds

process

correlations

y

t

= -0.067

y

t-1

-0.663

ϵ

t-1

+

ϵ

t

The autoregressive moving-average process (ARMA) is a discrete-time and continuous-state random process. This generator randomly chooses parameters of the model from the interval

(-1,1)

; you can set the condition for (weak) stationarity.

Part of the output includes the autocorrelation function (ACF), partial autocorrelation function (PACF), and their samples (SACF, SPACF), which serve as a basic tool for model identification in the Box–Jenkins approach by looking for so-called cut-off points.