neighbayes.dgp.simulate_sar_zinb¶
-
neighbayes.dgp.simulate_sar_zinb(n_side=
None, W=None, gdf=None, rho=0.5, rho_zero=0.3, beta=None, beta_zero=None, alpha=2.0, X_zero=None, X=None, W_zero=None, rng=None, seed=None, contiguity='queen', target_pi=None, err_hetero=False, create_gdf=False, geometry_type='polygon', latent=True, *, T=None, N=None, unit_effects=None, time_effects=False, unit_effect_sd=None, time_effect_sd=None)[source]¶ Simulate data from a zero-inflated SAR-NB DGP.
Two-part model:
Zero part (unit activity — reduced-form SAR logit):
\[d_i \sim \text{Bernoulli}(\text{logit}^{-1}(\eta_{z,i})), \quad \eta_z = (I - \rho_z W_z)^{-1} X_z\beta_z,\]the reduced-form zero part
SARZINBfits. (Until 2026-10-03 a latentν ~ N(0, I)was added inside the filter, a random effect the model does not have, which attenuated theβ_zestimates.)Count part (volume — reduced-form SAR-NB):
\[y_i \mid d_i = 1 \sim \text{NegBin}(\exp(\eta^{\text{cnt}}_i), \alpha), \quad \eta^{\text{cnt}} = (I - \rho W_{\text{cnt}})^{-1} X\beta\]Observation:
\[y_i = 0 \text{ if } d_i = 0, \quad y_i \sim \text{NegBin}(\cdot) \text{ if } d_i = 1\]With
latent=True, the default, it draws the latent model the unmarkedSARZINBfits, each part’s own disturbance propagated: \(d_i = \mathbf 1(u_i > 0)\), \(u = (I - \rho_z W_z)^{-1}(X_z\beta_z + \varepsilon_z)\) with \(\varepsilon_z\) logistic, and \(y_i \mid d_i = 1 \sim \mathrm{Poisson}(e^{\eta_i})\), \(\eta = (I - \rho W)^{-1}(X\beta + \varepsilon)\) with \(\varepsilon_i = \log\gamma_i\), \(\gamma_i \sim \mathrm{Gamma}(\alpha, \alpha)\);eta_zeroandeta_cntare then the drawnuandeta.- Parameters:¶
- n_side : int, optional¶
Side length of the square rook grid used when neither
Wnorgdfis supplied (n_side**2observations).- W : Graph or scipy.sparse matrix, optional¶
Spatial weights for the count part. Also used for the zero part when
W_zerois not provided.- gdf : GeoDataFrame, optional¶
Geodataframe used to construct weights.
- rho : float, default 0.5¶
Spatial autoregressive parameter for the count part.
- rho_zero : float, default 0.3¶
Spatial autoregressive parameter for the zero part.
- beta : ndarray, optional¶
Count-part coefficients (including intercept). Defaults to
[1.0, 0.6].- beta_zero : ndarray, optional¶
Zero-part coefficients (including intercept). Defaults to
[0.3, 1.0].- alpha : float, default 2.0¶
NB2 dispersion parameter. Must be strictly positive.
- X_zero : ndarray, optional¶
Zero-part covariate matrix of shape
(nobs, p). If not provided, a random design matrix is generated frombeta_zero.- X : ndarray, optional¶
Count covariate matrix of shape
(nobs, k). If not provided, a random design matrix is generated frombeta.- W_zero : Graph or scipy.sparse matrix, optional¶
Spatial weights for the zero part. If
None, usesW(the same weights in both parts).- rng : numpy.random.Generator, optional¶
Random number generator.
- seed : int, optional¶
Random seed (used only if rng is None).
- contiguity : str, default "queen"¶
Neighbor rule used when W is built from
gdf.- target_pi : float, optional¶
If given, an intercept shift is solved so that the marginal activation probability
mean(pi) == target_pi. The shift is added tobeta_zero[0](inside the filter, so the reportedbeta_zerogenerateseta_zero).- err_hetero : bool, default False¶
Not implemented for ZINB models. If True, a warning is issued and the parameter is ignored (homoskedastic errors are used).
- create_gdf : bool, default False¶
Whether to attach a GeoDataFrame to the output.
- geometry_type : str, default "polygon"¶
Type of geometry for the GeoDataFrame.
- Returns:¶
Keys:
y,d,X,X_zero,eta_cnt,eta_zero,W_sparse,W_graph,W_zero_sparse,W_zero_graph,params_true.- Return type:¶