Skip to content

ProphetGamma

Bases: Prophetverse

Gamma-likelihood prophet.

Parameters:

Name Type Description Default
changepoint_interval int

Number of potential changepoints to sample in the history.

25
changepoint_range float or int

Proportion of the history in which trend changepoints will be estimated.

  • if float, must be between 0 and 1. The range will be that proportion of the training history.

  • if int, ca nbe positive or negative. Absolute value must be less than number of training points. The range will be that number of points. A negative int indicates number of points counting from the end of the history, a positive int from the beginning.

0.8
changepoint_prior_scale float

Regularization parameter controlling the flexibility of the automatic changepoint selection.

0.001
offset_prior_scale float

Scale parameter for the prior distribution of the offset. The offset is the constant term in the piecewise trend equation.

0.1
feature_transformer sktime transformer, BaseTransformer

Transformer object to generate Fourier terms, holiday or other features. If None, no additional features are used. For multiple features, pass a FeatureUnion object with the transformers.

None
capacity_prior_scale float

Scale parameter for the prior distribution of the capacity.

0.2
capacity_prior_loc float

Location parameter for the prior distribution of the capacity.

1.1
noise_scale float

Scale parameter for the observation noise.

0.05
trend str, optional, one of "linear" (default) or "logistic"

Type of trend to use. Can be "linear" or "logistic".

'logistic'
mcmc_samples int

Number of MCMC samples to draw.

2000
mcmc_warmup int

Number of MCMC warmup steps. Also known as burn-in.

200
mcmc_chains int

Number of MCMC chains to run in parallel.

4
inference_method str, optional, one of "mcmc" or "map"

Inference method to use. Can be "mcmc" or "map".

"map"
optimizer_name str

Name of the numpyro optimizer to use for variational inference.

"Adam"
optimizer_kwargs dict

Additional keyword arguments to pass to the numpyro optimizer.

{}
optimizer_steps int

Number of optimization steps to perform for variational inference.

100_000
exogenous_effects List[AbstractEffect]

A list of prophetverse AbstractEffect objects defining the exogenous effects to be used in the model.

None
default_effect AbstractEffectm optional, defalut=None

The default effect to be used when no effect is specified for a variable.

None
default_exogenous_prior tuple

Default prior distribution for exogenous effects.

None
rng_key jax.random.PRNGKey or None (default

Random number generator key.

None
Source code in src/prophetverse/sktime/univariate.py
class ProphetGamma(Prophetverse):
    """Gamma-likelihood prophet.

    Parameters
    ----------
    changepoint_interval : int, optional, default=25
        Number of potential changepoints to sample in the history.

    changepoint_range : float or int, optional, default=0.8
        Proportion of the history in which trend changepoints will be estimated.

        * if float, must be between 0 and 1.
          The range will be that proportion of the training history.

        * if int, ca nbe positive or negative.
          Absolute value must be less than number of training points.
          The range will be that number of points.
          A negative int indicates number of points
          counting from the end of the history, a positive int from the beginning.

    changepoint_prior_scale : float, optional, default=0.001
        Regularization parameter controlling the flexibility
        of the automatic changepoint selection.

    offset_prior_scale : float, optional, default=0.1
        Scale parameter for the prior distribution of the offset.
        The offset is the constant term in the piecewise trend equation.

    feature_transformer : sktime transformer, BaseTransformer, optional, default=None
        Transformer object to generate Fourier terms, holiday or other features.
        If None, no additional features are used.
        For multiple features, pass a ``FeatureUnion`` object with the transformers.

    capacity_prior_scale : float, optional, default=0.2
        Scale parameter for the prior distribution of the capacity.

    capacity_prior_loc : float, optional, default=1.1
        Location parameter for the prior distribution of the capacity.

    noise_scale : float, optional, default=0.05
        Scale parameter for the observation noise.
    trend : str, optional, one of "linear" (default) or "logistic"
        Type of trend to use. Can be "linear" or "logistic".

    mcmc_samples : int, optional, default=2000
        Number of MCMC samples to draw.

    mcmc_warmup : int, optional, default=200
        Number of MCMC warmup steps. Also known as burn-in.

    mcmc_chains : int, optional, default=4
        Number of MCMC chains to run in parallel.

    inference_method : str, optional, one of "mcmc" or "map", default="map"
        Inference method to use. Can be "mcmc" or "map".

    optimizer_name : str, optional, default="Adam"
        Name of the numpyro optimizer to use for variational inference.

    optimizer_kwargs : dict, optional, default={}
        Additional keyword arguments to pass to the numpyro optimizer.

    optimizer_steps : int, optional, default=100_000
        Number of optimization steps to perform for variational inference.

    exogenous_effects : List[AbstractEffect], optional, default=None
        A list of ``prophetverse`` ``AbstractEffect`` objects
        defining the exogenous effects to be used in the model.

    default_effect : AbstractEffectm optional, defalut=None
        The default effect to be used when no effect is specified for a variable.

    default_exogenous_prior : tuple, default=None
        Default prior distribution for exogenous effects.

    rng_key : jax.random.PRNGKey or None (default
        Random number generator key.
    """

    def __init__(
        self,
        changepoint_interval=25,
        changepoint_range=0.8,
        changepoint_prior_scale=0.001,
        offset_prior_scale=0.1,
        feature_transformer=None,
        capacity_prior_scale=0.2,
        capacity_prior_loc=1.1,
        noise_scale=0.05,
        trend="logistic",
        mcmc_samples=2000,
        mcmc_warmup=200,
        mcmc_chains=4,
        inference_method="map",
        optimizer_name="Adam",
        optimizer_kwargs=None,
        optimizer_steps=100_000,
        exogenous_effects=None,
        default_effect=None,
        scale=None,
        rng_key=None,
    ):

        super().__init__(
            changepoint_interval=changepoint_interval,
            changepoint_range=changepoint_range,
            changepoint_prior_scale=changepoint_prior_scale,
            offset_prior_scale=offset_prior_scale,
            feature_transformer=feature_transformer,
            capacity_prior_scale=capacity_prior_scale,
            capacity_prior_loc=capacity_prior_loc,
            noise_scale=noise_scale,
            trend=trend,
            mcmc_samples=mcmc_samples,
            mcmc_warmup=mcmc_warmup,
            mcmc_chains=mcmc_chains,
            inference_method=inference_method,
            optimizer_name=optimizer_name,
            optimizer_kwargs=optimizer_kwargs,
            optimizer_steps=optimizer_steps,
            exogenous_effects=exogenous_effects,
            likelihood="gamma",
            default_effect=default_effect,
            scale=scale,
            rng_key=rng_key,
        )