col_gen_estimator.DTreeMasterProblem

class col_gen_estimator.DTreeMasterProblem(initial_paths, leaves, nodes, splits, beta_constraints_as_cuts=False, generate_cuts=False, num_cuts_round=0, solver_type='glop', data_rows=None)

Master problem for Decision Tree classifier.

__init__(initial_paths, leaves, nodes, splits, beta_constraints_as_cuts=False, generate_cuts=False, num_cuts_round=0, solver_type='glop', data_rows=None)

add_column(path)

TODO: Documentation.

create_cut_gen_model()

Creates the model for cut generation using a CP-SAT solver.

generate_mp(X, y)

TODO: Documentation.

get_satisfied_path_ids(row)

TODO: Documentation.

preprocess_rows(aggressive=False)

TODO: Documentation.

rmp_objective_improved()

(Optional) Returns True if objective value of rmp is improved. This is used for computing the number of improving iterations.

solve_ip(solver_params='')

Solves the integer RMP with given solver params. Returns True if the tree is generated.

Parameters:

solver_paramsstring, default=’’,

The solver parameters for solving the integer RMP.

solve_rmp(solver_params='')

TODO: Documentation.

store_lp_solution()

Store the solution values in the class.

violated_row_constraint(satisfied_path_ids)

TODO: Documentation.