neighbayes.models.priors.PanelGaussianPriors¶

class neighbayes.models.priors.PanelGaussianPriors(beta_mu=0.0, beta_sigma=1000000.0, sigma2_alpha=2.0, sigma2_beta=1.0, sigma2_y_alpha=2.0, sigma2_y_beta=1.0, gamma_prior_var=1.0, rho_lower=-0.999, rho_upper=0.999)[source]¶

Prior hyperparameters for the Gaussian panel flow Gibbs sampler.

All priors are weakly informative by default, matching the GibbsPriors / FlowGibbsPriors convention.

Parameters:¶

beta_mu : float, default 0.0¶: Normal prior mean for \(\beta\).
beta_sigma : float, default 1e6¶: Normal prior standard deviation for \(\beta\).
sigma2_alpha : float, default 2.0¶: Inverse-Gamma shape for \(\sigma^2_u\).
sigma2_beta : float, default 1.0¶: Inverse-Gamma scale for \(\sigma^2_u\).
sigma2_y_alpha : float, default 2.0¶: Inverse-Gamma shape for \(\sigma^2_y\).
sigma2_y_beta : float, default 1.0¶: Inverse-Gamma scale for \(\sigma^2_y\).
gamma_prior_var : float, default 1.0¶: Prior variance for \(\gamma \sim N(0, \sigma^2_\gamma)\) truncated to \((-1, 1)\).
rho_lower : float, default -0.999¶: Lower bound for \(\rho_d, \rho_o\).
rho_upper : float, default 0.999¶: Upper bound for \(\rho_d, \rho_o\).

__init__(beta_mu=0.0, beta_sigma=1000000.0, sigma2_alpha=2.0, sigma2_beta=1.0, sigma2_y_alpha=2.0, sigma2_y_beta=1.0, gamma_prior_var=1.0, rho_lower=-0.999, rho_upper=0.999)[source]¶

Methods

`__init__`([beta_mu, beta_sigma, ...])

Attributes