arch.univariate.base.ARCHModelForecast¶

class arch.univariate.base.ARCHModelForecast(index: list[Hashable | None] | Index, start_index: int, mean: ndarray[tuple[int, ...], dtype[float64]], variance: ndarray[tuple[int, ...], dtype[float64]], residual_variance: ndarray[tuple[int, ...], dtype[float64]], simulated_paths: ndarray[tuple[int, ...], dtype[float64]] | None = None, simulated_variances: ndarray[tuple[int, ...], dtype[float64]] | None = None, simulated_residual_variances: ndarray[tuple[int, ...], dtype[float64]] | None = None, simulated_residuals: ndarray[tuple[int, ...], dtype[float64]] | None = None, align: 'origin' | 'target' | 'horizon' = 'origin', *, reindex: bool = False)[source]¶

Container for forecasts from an ARCH Model.

Parameters:¶

index: list[Hashable | None] | Index¶: The index of the data used to produce the forecasts.
start_index: int¶: The position in index of the first row of the forecasts.
mean: ndarray[tuple[int, ...], dtype[float64]]¶: Forecasts of the conditional mean. A 2-d array with one row for each forecast origin and one column for each step of the forecast horizon.
variance: ndarray[tuple[int, ...], dtype[float64]]¶: Forecasts of the conditional variance, with the same dimensions as mean.
residual_variance: ndarray[tuple[int, ...], dtype[float64]]¶: Forecasts of the conditional variance of the residuals, with the same dimensions as mean.
simulated_paths: ndarray[tuple[int, ...], dtype[float64]] | None = None¶: Simulated paths of the process. A 3-d array with dimensions (number of forecast origins, number of simulations, horizon).
simulated_variances: ndarray[tuple[int, ...], dtype[float64]] | None = None¶: Simulated variances of the process, with the same dimensions as simulated_paths.
simulated_residual_variances: ndarray[tuple[int, ...], dtype[float64]] | None = None¶: Simulated variances of the residuals, with the same dimensions as simulated_paths.
simulated_residuals: ndarray[tuple[int, ...], dtype[float64]] | None = None¶: Simulated residuals, with the same dimensions as simulated_paths.
align: 'origin' | 'target' | 'horizon' = 'origin'¶: Alignment of the forecasts. See the align parameter of the forecast method. ‘horizon’ is an alias of ‘target’.
reindex: bool = False¶: Whether to reindex the forecasts to have the same dimension as the series being forecast.

Notes

The container is created by the forecasting methods. The arguments of the constructor are not validated.

Methods

Properties

`mean`	Forecast values for the conditional mean of the process.
`residual_variance`	Forecast values for the conditional variance of the residuals.
`simulations`	Detailed simulation results if using a simulation-based method.
`variance`	Forecast values for the conditional variance of the process.