New study reveals key to more accurate future predictions!
The random walk model with drift can predict future values of a specific type of process called ARIMA (0,2,1). When the parameter θ is close to –1, the random walk model behaves like the ARIMA (0,2,1). For values of θ not so close to –1, the prediction intervals may not be accurate. However, when forecasting values with θ close to –1, the random walk model is better, producing narrower confidence intervals with similar coverage to the ARIMA (0,2,1).
