TARNNResult._from_fit#

classmethod TARNNResult._from_fit(fit, model)[source]#

Assemble the public result from a raw fit and its specification.

The threshold variable is re-lagged here from the series or the external variable so that the stored copy is aligned with the effective sample rather than the full one.

Parameters:

  • fit (_NeuralThresholdFit) – The raw estimator output.

  • model (_NeuralThresholdModel[TARNNResult]) – The specification, read for the series, the order and the engine.

Return type:

TARNNResult

Returns:

The assembled result.

Example

>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> model = TARNN(rng.standard_normal(300), order=1, delay=2)
>>> res = TARNNResult._from_fit(model._fit_family(), model)
>>> res.delay, res.nobs, res.threshold_values.shape
(2, 298, (298,))