col_gen_estimator.DTreeSubProblem

class col_gen_estimator.DTreeSubProblem(leaf, nodes, splits, targets, depth, data_rows=None, solver_str='sat')

TODO: Documentation.

__init__(leaf, nodes, splits, targets, depth, data_rows=None, solver_str='sat') → None

create_submip(leaf_dual, row_duals, ns_duals, cut_duals)

TODO: Documentation.

generate_columns(X, y, dual_costs, params='')

TODO: Documentation.

row_satisfies_split(row, split)

TODO: Documentation.

update_objective(leaf_dual, row_duals, ns_duals, cut_duals)

TODO: Documentation.

update_subproblem(X, y, dual_costs)

(Optional) Updates the subproblem model. This is useful when we add more constraints to the master problem. The subproblem needs to be updated to handle new dual cost information. Parameters ———- X : ndarray, shape (n_samples, n_features)

The input.

yndarray, shape (n_samples,)

The labels.

dual_costs:

The dual costs and other information needed to update the subproblem.