ARMA

For forecasting time series that are stationary (see stationarity (stochastic process)).

Autoregressive moving average – the variable of interest is a linear combination of past values. The error (regression) is a linear combination of past values.