causaldag.utils.scores.local_gaussian_bge_score

causaldag.utils.scores.local_gaussian_bge_score(node, parents, suffstat, alpha_mu=None, alpha_w=None, inverse_scale_matrix=None, parameter_mean=None, is_diagonal=True)

Compute the BGE score of a node given its parents.

Parameters:

• node – TODO - describe.
• parents – TODO - describe.
• suffstat –

  dictionary containing:

  • n – number of samples
  • S – sample covariance matrix
  • mu – sample mean
• alpha_mu – TODO - describe. Default is the number of variables.
• alpha_w – TODO - describe. Default is the (number of variables) + alpha_mu + 1
• inverse_scale_matrix – TODO - describe. Default is the identity matrix.
• parameter_mean – TODO - describe. Default is the zero vector.
• is_diagonal – TODO - describe.

Returns:

BGE score.

Return type:

float