causalis.dgp.panel_data_scm.base¶
Module Contents¶
Classes¶
Low-level panel SCM generator supporting Gaussian, Gamma, and Poisson outcomes. |
API¶
- class causalis.dgp.panel_data_scm.base.PanelSCMGeneratorConfig¶
- n_donors: int¶
5
- n_pre_periods: int¶
20
- n_post_periods: int¶
10
- time_start: int¶
1
- time_freq: str¶
‘M’
- calendar_start: str¶
‘2000-01’
- treated_unit: Hashable¶
‘treated’
- donor_prefix: str¶
‘donor_’
- random_state: Optional[int]¶
42
- return_panel_data: bool¶
True
- dirichlet_alpha: float¶
1.0
- rho_common: float¶
0.0
- rho_donor: float¶
0.0
- n_latent_factors: int¶
0
- latent_loading_std: float¶
0.35
- rho_latent: float¶
0.0
- rho_prefit_mismatch: float¶
0.0
- outcome_distribution: Literal[gaussian, gamma, poisson]¶
‘gaussian’
- treatment_effect: float¶
2.0
- treatment_effect_slope: float¶
0.0
- donor_noise_std: float¶
0.2
- treated_noise_std: float¶
0.1
- common_factor_std: float¶
0.15
- latent_factor_std: float¶
0.2
- prefit_mismatch_std: float¶
0.0
- treatment_effect_mode: Literal[additive, multiplicative]¶
‘additive’
- treatment_effect_rate: float¶
0.12
- gamma_shape: float¶
6.0
- donor_noise_std_log: float¶
0.15
- common_factor_std_log: float¶
0.1
- latent_factor_std_log: float¶
0.1
- prefit_mismatch_std_log: float¶
0.08
- class causalis.dgp.panel_data_scm.base.PanelSCMGenerator(config: causalis.dgp.panel_data_scm.base.PanelSCMGeneratorConfig)¶
Low-level panel SCM generator supporting Gaussian, Gamma, and Poisson outcomes.
Initialization
- generate(*, return_panel_data: Optional[bool] = None) Union[pandas.DataFrame, causalis.data_contracts.panel_data_scm.PanelDataSCM]¶