statsmodels.tsa.vector_ar.var_model.VARProcess.forecast_interval

VARProcess.forecast_interval(y, steps, alpha=0.05, exog_future=None)[source]

Construct forecast interval estimates assuming the y are Gaussian

Parameters:

  • y ({ndarray, None}) – The initial values to use for the forecasts. If None, the last k_ar values of the original endogenous variables are used.

  • steps (int) – Number of steps ahead to forecast

  • alpha (float, optional) – The significance level for the confidence intervals.

  • exog_future (ndarray, optional) – Forecast values of the exogenous variables. Should include constant, trend, etc. as needed, including extrapolating out of sample.

Returns:

  • point (ndarray) – Mean value of forecast

  • lower (ndarray) – Lower bound of confidence interval

  • upper (ndarray) – Upper bound of confidence interval

Notes

Lütkepohl pp. 39-40