Applying ARIMA Forecasts
by ERIC WEISS, Ph.D.
"Applying ARIMA Forecasts" is an article continuing a series started in the October (82) and the
January (83) issues of Technical Analysis. AutoRegressive-lntegrated Moving Average(ARIMA) is a
forecasting methodology based upon the techniques described by Box and Jenkins in their book: Time
Series Analysis, Forecasting and Control, published by Holden-Day, 1976, 2nd ed.
With understanding and time one may construct a model of a discrete statistical time series, such as the
markets, that will forecast the future with known probable error.
To begin, let's use this example:
Sum of Squared Residuals: 44.76
Residual Variance: .75
Average Absolute Error: .65
Degrees of Freedom: 60