INNER CODE UNIT · Python

model_training_order

AnotherSamWilson/miceforest · miceforest/imputation_kernel.py:201

        model_training_order = self.imputation_order + modeled_but_not_imputed_variables
        self.model_training_order = model_training_order

        self.initialize_empty = initialize_empty
        self.save_all_iterations_data = save_all_iterations_data

        # Models are stored in a dict, keys are (variable, iteration, dataset)
        self.models: Dict[Tuple[str, int, int], Booster] = {}

        # Candidate preds are stored the same as models.
        self.candidate_preds: Dict[str, DataFrame] = {}

        # Optimal parameters can only be found on 1 dataset at the current iteration.
        self.optimal_parameters: Dict[str, Dict[str, Any]] = {}

        # Determine available candidates and interpret data subset.
        available_candidates = {
            v: (self.shape[0] - self.na_counts[v]) for v in self.model_training_order

View source record →