A nonseasonal ARIMA model is classified as an “ARIMA(p,d,q)” model, where: p is the number of autoregressive terms, d is the number of nonseasonal differences needed for stationarity, and . q is the number of lagged forecast errors in the prediction equation .