smooth.ADAM.distribution_

property ADAM.distribution_: str

$distribution).

The probability distribution assumed for the error term in the model. Uses trailing underscore to distinguish from the input parameter and follow scikit-learn convention for fitted attributes.

Returns:

Distribution name, one of: - "dnorm": Normal distribution (default for additive errors) - "dgamma": Gamma distribution (default for multiplicative) - "dlaplace": Laplace distribution - "dlnorm": Log-Normal distribution - "dinvgauss": Inverse Gaussian distribution - "ds": S distribution - "dgnorm": Generalized Normal distribution

Return type:

str

Raises:

ValueError – If the model has not been fitted yet.

See also

loss_

Loss function used for estimation

Examples

>>> model = ADAM(model="AAN", distribution="dlaplace")
>>> model.fit(y)
>>> print(model.distribution_)
'dlaplace'

Type:

Error distribution used for fitting (R