MixedLogit.preprocess_model_data

MixedLogit.preprocess_model_data(choice_df, utility_equations)

Pre-process the model initiation inputs into PyMC-ready format.

This method prepares: - X: 3D design matrix (n_obs, n_alts, n_covariates) - F: Fixed covariate matrix (n_obs, n_fixed_covariates) or None - y: Encoded response vector (n_obs,) - group_idx: Group membership array (n_obs,) or None

Also extracts and stores metadata including alternatives, covariate names, and coordinate dimensions.

Parameters:

choice_dfpd.DataFrame

DataFrame containing choice observations and covariates

utility_equationslist[str]

List of utility formulas, one per alternative

Returns:

tuple

(X, F, y) ready for model building