Autoregressive integrated moving average (ARIMA) is a statistical model that is used to forecast future values of a time series. 

It is a generalization of the autoregressive moving average (ARMA) model, which only considers the autocorrelations between the current value and the past values of the time series. 

ARIMA also considers the moving average of the residuals of the ARMA model, which helps to improve the accuracy of the forecasts.

In simple words, ARIMA models the time series as a combination of three components:

• Autoregression (AR): The AR component models the dependence between the current value and the past values of the time series.

• Moving average (MA): The MA component models the dependence between the current value and the errors of the previous predictions.

• Integration (I): The I component models the non-stationarity of the time series by differencing the data.

ARIMA models are commonly used to forecast financial data, such as stock prices and exchange rates. They are also used to forecast other types of time series data, such as customer demand, sales, and inventory levels.