ARIMA

arima.ARIMA · inherits BaseForecaster

Represents a fixed-order ARIMA forecaster that standardizes training data for optimization stability and returns forecasts in the original value scale.

Attributes: * uses_exog: True * alias: Display name for external reporting. * model_: dict[str, Any] | None. Fitted model payload after fit.

__init__(self, order=(0, 0, 0), seasonal_order=(0, 0, 0), period=1, include_mean=True, method='CSS', alias='ARIMA', standardize=True)

Initialize a fixed-order ARIMA estimator.

Parameter	Type	Default	Description
order	tuple[int, int, int]	(0, 0, 0)	Non-seasonal order (p, d, q).
seasonal_order	tuple[int, int, int]	(0, 0, 0)	Seasonal order (P, D, Q).
period	int	1	Seasonal cycle length.
include_mean	bool	True	Whether to include deterministic mean/drift.
method	str	"CSS"	Objective method strategy.
alias	str	"ARIMA"	User-facing model name.
standardize	bool	True	Enables series standardization.

fit(self, y, X=None) -> Self

Fit ARIMA parameters on a training series.

Parameters:

Parameter	Type	Default	Description
y	jnp.ndarray	-	Training target series.
X	Optional[jnp.ndarray]	None	Optional exogenous matrix aligned with y.

Returns: Self (The fitted estimator instance; sets self.model_). Raises: RuntimeError (Propagated from fitting internals if optimization fails irrecoverably.)

forecast(self, h, y, X=None, X_future=None, level=None, fitted=False) -> dict[str, jnp.ndarray]

Fit and forecast in one call.

Parameters:

Parameter	Type	Default	Description
h	int	-	Forecast horizon.
y	jnp.ndarray	-	Source training series.
X	Optional[jnp.ndarray]	None	Exogenous regressors. The fast path currently does not use exogenous variables.
X_future	Optional[jnp.ndarray]	None	Future exogenous regressors (unused; included for BaseForecaster compliance).
level	Optional[list]	None	Confidence levels (unused; included for BaseForecaster compliance).
fitted	bool	False	Whether to return fitted values (unused; included for BaseForecaster compliance).

Returns: dict[str, jnp.ndarray] (Forecast dictionary containing mean). Raises: RuntimeError (Propagated if optimizer objective diverges.)

predict(self, h, X=None, level=None) -> dict[str, jnp.ndarray]

Generate forecasts from a fitted ARIMA model.

Parameters:

Parameter	Type	Default	Description
h	int	-	Number of future steps to predict.
X	Optional[jnp.ndarray]	None	Optional exogenous future matrix.
level	int \| tuple[int, ...] \| None	None	Confidence levels for interval generation.

Returns: dict[str, jnp.ndarray] (Dictionary containing mean and optional lower/upper interval keys (e.g., lo-95, hi-95).) Raises: RuntimeError (If called before fit.)