Skip to content

Commit

Permalink
Fix a_scale, b_scale dims
Browse files Browse the repository at this point in the history
  • Loading branch information
@PabloRoque
PabloRoque committed Feb 6, 2025
1 parent 60372e0 commit 96c2609
Show file tree
Hide file tree
Showing 2 changed files with 29 additions and 36 deletions.
59 changes: 26 additions & 33 deletions pymc_marketing/clv/models/beta_geo.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,7 @@ class BetaGeoModel(CLVModel):
r"""Beta-Geometric Negative Binomial Distribution (BG/NBD) model for a non-contractual customer population across continuous time.
First introduced by Fader, Hardie & Lee [1]_, with additional predictive methods
and enhancements in [2]_,[3]_ and [4]_.
and enhancements in [2]_,[3]_, [4]_ and [5]_
The BG/NBD model assumes dropout probabilities for the customer population are Beta distributed,
and time between transactions follows a Gamma distribution while the customer is still active.
Expand Down Expand Up @@ -214,9 +214,9 @@ def build_model(self) -> None: # type: ignore[override]
self.model_config[
"purchase_coefficient_prior"
].dims = "purchase_covariate"
purchase_coefficient_gamma1 = self.model_config[
purchase_coefficient_alpha = self.model_config[
"purchase_coefficient_prior"
].create_variable("purchase_coefficient_gamma1")
].create_variable("purchase_coefficient_alpha")

alpha_scale = self.model_config["alpha_prior"].create_variable(
"alpha_scale"
Expand All @@ -226,7 +226,7 @@ def build_model(self) -> None: # type: ignore[override]
(
alpha_scale
* pm.math.exp(
-pm.math.dot(purchase_data, purchase_coefficient_gamma1)
-pm.math.dot(purchase_data, purchase_coefficient_alpha)
)
),
dims="customer_id",
Expand All @@ -246,29 +246,25 @@ def build_model(self) -> None: # type: ignore[override]
self.model_config[
"dropout_coefficient_prior"
].dims = "dropout_covariate"
dropout_coefficient_gamma2 = self.model_config[
dropout_coefficient_a = self.model_config[
"dropout_coefficient_prior"
].create_variable("dropout_coefficient_gamma2")
dropout_coefficient_gamma3 = self.model_config[
].create_variable("dropout_coefficient_a")
dropout_coefficient_b = self.model_config[
"dropout_coefficient_prior"
].create_variable("dropout_coefficient_gamma3")
].create_variable("dropout_coefficient_b")

a_scale = self.model_config["a_prior"].create_variable("a_scale")
b_scale = self.model_config["b_prior"].create_variable("b_scale")
a = pm.Deterministic(
"a",
a_scale
* pm.math.exp(
pm.math.dot(dropout_data, dropout_coefficient_gamma2)
),
* pm.math.exp(pm.math.dot(dropout_data, dropout_coefficient_a)),
dims="customer_id",
)
b = pm.Deterministic(
"b",
b_scale
* pm.math.exp(
pm.math.dot(dropout_data, dropout_coefficient_gamma3)
),
* pm.math.exp(pm.math.dot(dropout_data, dropout_coefficient_b)),
dims="customer_id",
)
else:
Expand All @@ -286,43 +282,40 @@ def build_model(self) -> None: # type: ignore[override]
self.model_config[
"dropout_coefficient_prior"
].dims = "dropout_covariate"
dropout_coefficient_gamma2 = self.model_config[
dropout_coefficient_a = self.model_config[
"dropout_coefficient_prior"
].create_variable("dropout_coefficient_gamma2")
dropout_coefficient_gamma3 = self.model_config[
].create_variable("dropout_coefficient_a")
dropout_coefficient_b = self.model_config[
"dropout_coefficient_prior"
].create_variable("dropout_coefficient_gamma3")
].create_variable("dropout_coefficient_b")

phi_dropout = self.model_config[
"phi_dropout_prior"
].create_variable("phi_dropout")

kappa_dropout = self.model_config[
"kappa_dropout_prior"
].create_variable("kappa_dropout")

a_scale = pm.Deterministic(
"a_scale", phi_dropout * kappa_dropout, dims="customer_id"
"a_scale",
phi_dropout * kappa_dropout,
)
b_scale = pm.Deterministic(
"b_scale",
(1.0 - phi_dropout) * kappa_dropout,
dims="customer_id",
)

a = pm.Deterministic(
"a",
a_scale
* pm.math.exp(
pm.math.dot(dropout_data, dropout_coefficient_gamma2)
),
* pm.math.exp(pm.math.dot(dropout_data, dropout_coefficient_a)),
dims="customer_id",
)
b = pm.Deterministic(
"b",
b_scale
* pm.math.exp(
pm.math.dot(dropout_data, dropout_coefficient_gamma3)
),
* pm.math.exp(pm.math.dot(dropout_data, dropout_coefficient_b)),
dims="customer_id",
)

Expand Down Expand Up @@ -393,11 +386,11 @@ def _extract_predictive_variables(
coords=[customer_id, list(model_coords["purchase_covariate"])],
)
alpha_scale = self.fit_result["alpha_scale"]
purchase_coefficient_gamma1 = self.fit_result["purchase_coefficient_gamma1"]
purchase_coefficient_alpha = self.fit_result["purchase_coefficient_alpha"]
alpha = alpha_scale * np.exp(
-xarray.dot(
purchase_xarray,
purchase_coefficient_gamma1,
purchase_coefficient_alpha,
dim="purchase_covariate",
)
)
Expand All @@ -412,19 +405,19 @@ def _extract_predictive_variables(
coords=[customer_id, list(model_coords["dropout_covariate"])],
)
a_scale = self.fit_result["a_scale"]
dropout_coefficient_gamma2 = self.fit_result["dropout_coefficient_gamma2"]
dropout_coefficient_gamma3 = self.fit_result["dropout_coefficient_gamma3"]

dropout_coefficient_a = self.fit_result["dropout_coefficient_a"]
a = a_scale * np.exp(
xarray.dot(
dropout_xarray, dropout_coefficient_gamma2, dim="dropout_covariate"
dropout_xarray, dropout_coefficient_a, dim="dropout_covariate"
)
)
a.name = "a"

dropout_coefficient_b = self.fit_result["dropout_coefficient_b"]
b_scale = self.fit_result["b_scale"]
b = b_scale * np.exp(
xarray.dot(
dropout_xarray, dropout_coefficient_gamma3, dim="dropout_covariate"
dropout_xarray, dropout_coefficient_b, dim="dropout_covariate"
)
)
b.name = "b"
Expand Down
6 changes: 3 additions & 3 deletions pymc_marketing/clv/utils.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -161,12 +161,12 @@ def _find_first_transactions(
customer_id_col: str,
datetime_col: str,
monetary_value_col: str | None = None,
purchase_covariate_cols: list | None = None,
dropout_covariate_cols: list | None = None,
datetime_format: str | None = None,
observation_period_end: str | pandas.Period | datetime | None = None,
time_unit: str = "D",
sort_transactions: bool | None = True,
purchase_covariate_cols: list | None = None,
dropout_covariate_cols: list | None = None,
) -> pandas.DataFrame:
"""Return dataframe with first transactions.
Expand Down Expand Up @@ -225,7 +225,7 @@ def _find_first_transactions(
[
extra_col
for extra_col in dropout_covariate_cols
if extra_col not in purchase_covariate_cols
if extra_col not in purchase_covariate_cols # type: ignore
]
)

Expand Down

0 comments on commit 96c2609

Please sign in to comment.