HSGPKwargs#

pydantic model pymc_marketing.hsgp_kwargs.HSGPKwargs[source]#

HSGP keyword arguments for the time-varying prior.

See [1] and [2] for the theoretical background on the Hilbert Space Gaussian Process (HSGP). See , [6] for a practical guide through the method using code examples. See the HSGP class for more information on the Hilbert Space Gaussian Process in PyMC. We also recommend the following resources for a more practical introduction to HSGP: [3], [4], [5].

References

[1]

Solin, A., Sarkka, S. (2019) Hilbert Space Methods for Reduced-Rank Gaussian Process Regression.

[2]

Ruitort-Mayol, G., and Anderson, M., and Solin, A., and Vehtari, A. (2022). Practical Hilbert Space Approximate Bayesian Gaussian Processes for Probabilistic Programming.

[5]

PyMC Example Gallery: “Baby Births Modelling with HSGPs”.

Methods

HSGPKwargs.__init__(**data)

Create a new model by parsing and validating input data from keyword arguments.

HSGPKwargs.from_dict(data)

Reconstruct from a dict.

HSGPKwargs.to_dict([_orig])

field L: float | None = None[source]#

Extent of basis functions. Set this to reflect the expected range of in+out-of-sample data (considering that time-indices are zero-centered). Defaults to X_mid * 2 (identical to c=2 in HSGP).

Constraints:

  • gt = 0

field cov_func: CovFunc | None = None[source]#

Covariance function enum. Supported values: ExpQuad, Matern52, Matern32. None is resolved to Matern52 at model-build time.

field eta_lam: float = 1.0[source]#

Exponential prior for the variance.

Constraints:

  • gt = 0

field ls_mu: float = 5.0[source]#

Mean of the inverse gamma prior for the lengthscale.

Constraints:

  • gt = 0

field ls_sigma: float = 5.0[source]#

Standard deviation of the inverse gamma prior for the lengthscale.

Constraints:

  • gt = 0

field m: int = 200[source]#

Number of basis functions.