model_arima

model_arima(
    data,
    p=1,
    d=0,
    q=1,
    include_intercept=True,
    transform=None,
    training_time_steps=None,
    parameters=None,
    parameter_values=None,
    use_default_flat_priors=None,
)

An ARIMA(p, d, q) model.

Parameters

Name	Type	Description	Default
data	array_like	The observed series, in sequence order.	required
p	int	Autoregressive, differencing, and moving-average orders.	`1`
d	int	Autoregressive, differencing, and moving-average orders.	`1`
q	int	Autoregressive, differencing, and moving-average orders.	`1`
include_intercept	bool	Include an intercept term.	`True`
transform	('None', 'Logarithmic', 'BoxCox', 'YeoJohnson')	Variance-stabilizing transform.	`"None"`
training_time_steps	int	Leading steps used for calibration, the rest held back for validation. The model default is `max(30, floor(0.8 * n))`, which exceeds the series length for any series shorter than 30 and then fails :func:`model_validate`; set it explicitly for a short series.	`None`
parameters		As in :func:`model_univariate`.	`None`
parameter_values		As in :func:`model_univariate`.	`None`
use_default_flat_priors		As in :func:`model_univariate`.	`None`

Returns

Name	Type	Description
	Model	The assembled model spec.