GammaGammaModelIndividual

class pymc_marketing.clv.models.gamma_gamma.GammaGammaModelIndividual(data, model_config=None, sampler_config=None)

Gamma-Gamma Model for expected future monetary value.

The Gamma-Gamma model assumes expected future spend follows a Gamma distribution, and the scale of this distribution is also Gamma-distributed.

This model is conditioned on the spend values of each purchase for each customer, and is based on [1], [2].

See GammaGammaModel for an equivalent model conditioned on mean transaction values of repeat purchases for the customer population.

Parameters:

dataDataFrame

Dataframe containing the following columns:

• customer_id: Unique customer identifier

• individual_transaction_value: Monetary value of each purchase for each customer

model_configdict, optional

Dictionary of model prior parameters. If not provided, the model will use default priors specified in the default_model_config class attribute.

sampler_configdict, optional

Dictionary of sampler parameters. Defaults to None.

References

[1]

Fader, P. S., & Hardie, B. G. (2013). “The Gamma-Gamma model of monetary value”. http://www.brucehardie.com/notes/025/gamma_gamma.pdf

[2]

Peter S. Fader, Bruce G. S. Hardie, and Ka Lok Lee (2005), “RFM and CLV: Using iso-value curves for customer base analysis”, Journal of Marketing Research, 42 (November), 415-430. https://journals.sagepub.com/doi/pdf/10.1509/jmkr.2005.42.4.415

Examples

import pymc as pm
from pymc_marketing.clv import GammaGammaModelIndividual

model = GammaGammaModelIndividual(
    data=pandas.DataFrame(
        {
        "customer_id": [0, 0, 0, 1, 1, 2, ...],
        "individual_transaction_value": [5.3. 5.7, 6.9, 13.5, 0.3, 19.2 ...],
        }
    ),
    model_config={
        "p": {dist: 'HalfNorm', kwargs: {}},
        "q": {dist: 'HalfStudentT', kwargs: {"nu": 4, "sigma": 10}},
        "v": {dist: 'HalfCauchy', kwargs: {}},
    },
    sampler_config={
        "draws": 1000,
        "tune": 1000,
        "chains": 2,
        "cores": 2,
        "nuts_kwargs": {"target_accept": 0.95},
    },
)

model.fit()
print(model.fit_summary())

# Predict spend of customers for which we know transaction history,
# conditioned on data. May include customers not included in fitting
expected_customer_spend = model.expected_customer_spend(
    data=pandas.DataFrame(
        {
        "customer_id": [0, 0, 0, 1, 1, 2, ...],
        "individual_transaction_value": [5.3. 5.7, 6.9, 13.5, 0.3, 19.2 ...],
        }
    ),
)
print(expected_customer_spend.mean("customer_id"))

# Predict spend of 10 new customers, conditioned on data
new_customer_spend = model.expected_new_customer_spend(n=10)
print(new_customer_spend.mean("new_customer_id"))

Methods

GammaGammaModelIndividual.__init__(data[, ...])	Initialize model configuration and sampler configuration for the model.
GammaGammaModelIndividual.attrs_to_init_kwargs(attrs)	Convert the model configuration and sampler configuration from the attributes to keyword arguments.
GammaGammaModelIndividual.build_from_idata(idata)	Build model from the InferenceData object.
GammaGammaModelIndividual.build_model()	Build the model.
GammaGammaModelIndividual.create_fit_data(X, y)	Create the fit_data group based on the input data.
GammaGammaModelIndividual.create_idata_attrs()	Create attributes for the inference data.
GammaGammaModelIndividual.distribution_customer_spend(data)	Posterior distribution of mean spend values for each customer.
GammaGammaModelIndividual.distribution_new_customer_spend([...])	Posterior distribution of mean spend values for new customers.
GammaGammaModelIndividual.expected_customer_lifetime_value(...)	Compute the average lifetime value for a group of one or more customers.
GammaGammaModelIndividual.expected_customer_spend(data)	Compute the expected future mean spend value per customer.
GammaGammaModelIndividual.expected_new_customer_spend()	Compute the expected mean spend value for a new customer.
GammaGammaModelIndividual.fit([method, ...])	Infer model posterior.
GammaGammaModelIndividual.fit_summary(**kwargs)	Compute the summary of the fit result.
GammaGammaModelIndividual.graphviz(**kwargs)	Get the graphviz representation of the model.
GammaGammaModelIndividual.load(fname)	Create a ModelBuilder instance from a file.
GammaGammaModelIndividual.load_from_idata(idata)	Create a ModelBuilder instance from an InferenceData object.
GammaGammaModelIndividual.post_sample_model_transformation()	Perform transformation on the model after sampling.
GammaGammaModelIndividual.predict([X, ...])	Use a model to predict on unseen data and return point prediction of all the samples.
GammaGammaModelIndividual.predict_posterior([...])	Generate posterior predictive samples on unseen data.
GammaGammaModelIndividual.predict_proba([X, ...])	Alias for predict_posterior, for consistency with scikit-learn probabilistic estimators.
GammaGammaModelIndividual.sample_posterior_predictive([...])	Sample from the model's posterior predictive distribution.
GammaGammaModelIndividual.sample_prior_predictive([...])	Sample from the model's prior predictive distribution.
GammaGammaModelIndividual.save(fname)	Save the model's inference data to a file.
GammaGammaModelIndividual.set_idata_attrs([idata])	Set attributes on an InferenceData object.
GammaGammaModelIndividual.thin_fit_result(...)	Return a copy of the model with a thinned fit result.

Attributes

X
default_model_config	Default model configuration.
default_sampler_config	Default sampler configuration.
fit_result	Get the posterior fit_result.
id	Generate a unique hash value for the model.
output_var	Output variable of the model.
posterior
posterior_predictive
predictions
prior
prior_predictive
version
y