Dot#

class pymc_marketing.terms.Dot(*, var_name, prior, name=None)[source]#

Linear predictor term: data @ beta.

Registers the data variable as pmd.Data for out-of-sample prediction.

Parameters:

var_namestr: Name of the data variable in the dataset.
priorVariableFactory: Prior for the beta coefficients. Any VariableFactory is accepted. Must have dims matching the last dimension of the data variable.
namestr, optional: Name for the coefficient variable. Defaults to {var_name}_beta. Provide a distinct name to reference the same var_name from multiple terms (e.g. separate alpha and beta coefficient branches) without colliding on the coefficient variable name.

Examples

from pymc_extras.prior import Prior
from pymc_marketing.terms import (
    Dot,
    Intercept,
    collect_coords,
    register_data,
    build_param,
)

# basic
mu = Intercept("mu") + Dot(var_name="x", prior=Prior("Normal", dims="feature"))

# two branches on the same data, distinct coefficient names
alpha_branch = Dot(
    var_name="x", name="alpha_coef", prior=Prior("Normal", dims="feature")
)
beta_branch = Dot(
    var_name="x", name="beta_coef", prior=Prior("Normal", dims="feature")
)

Methods

`Dot.__init__`(*, var_name, prior[, name])
`Dot.add_coords`(ds)	Add coordinates to the model during `register_data`.
`Dot.create_variable`()	Build `data @ beta` tensor.
`Dot.from_dict`(data)	Reconstruct a dot term from its serialized form.
`Dot.get_coords`(ds)	Extract coordinates from the data variable.
`Dot.register_data`(ds)	Register the data variable as `pmd.Data`.
`Dot.set_data`(ds[, model])	Update `pmd.Data` for prediction.
`Dot.to_dict`([_orig])

Attributes

`Dot.name`
`Dot.var_name`
`Dot.prior`