smooth.ADAM.predict_intervals

ADAM.predict_intervals(h, X=None, levels=[0.8, 0.95], side='both', nsim=10000)

Generate prediction intervals using the fitted ADAM model.

Convenience wrapper around predict() that defaults to interval="prediction" and accepts multiple confidence levels.

Parameters:

  • h (int) – Forecast horizon (number of steps to forecast).

  • X (Optional[NDArray], default=None) – Exogenous variables for the forecast period.

  • levels (List[float], default=[0.8, 0.95]) – Confidence levels for prediction intervals. Each level produces a pair of lower/upper columns in the output DataFrame. For example, levels=[0.8, 0.95] with side="both" yields columns "lower_0.1", "lower_0.025", "upper_0.9", "upper_0.975".

  • side (Literal["both", "upper", "lower"], default="both") – Which side(s) of the intervals to return.

  • nsim (int, default=10000) – Number of simulations for simulation-based intervals.

Returns:

DataFrame with "mean" and lower/upper columns for each level.

Return type:

pd.DataFrame

Parent Class: ADAM