Skip to main content

Table 1 Tips for selecting most appropriate autoregressive (p) and moving average (q) terms from autocorrelation and partial autocorrelation

From: Interrupted time series analysis using autoregressive integrated moving average (ARIMA) models: a guide for evaluating large-scale health interventions

Model type

Characteristics of ACF and PACF

ACF

PACF

ARIMA(p,d,0)

Tails off or is sinusoidal

Cuts off lag p

ARIMA(0,d,q)

Cuts off lag q

Tails off or is sinusoidal

ARMA(p,d,q)

Tails off or is sinusoidal

Tails off or is sinusoidal