paradance.optimization.MultipleObjective

class paradance.optimization.MultipleObjective(calculator: Calculator | LogarithmPCACalculator, direction: str | None = None, formula: str | None = None, warmup_formula: str | None = None, warmup_trials: int = 200, first_order: bool | None = False, power: bool | None = True, dirichlet: bool | None = False, weights_num: int | None = None, study_name: str | None = None, study_path: str | None = None, save_study: bool | None = True, checkpoint_path: str | None = None, first_order_with_scales: bool = True, first_order_lower_bound: float = 0.001, first_order_upper_bound: float = 1000000.0, free_style_lower_bound: float | List[float] = 0.001, free_style_upper_bound: float | List[float] = 1000000.0, base_weights: List[float] | None = None, base_weights_offset_ratio: float = 0.1, max_min_scale_ratio: float | None = None, first_order_scale_upper_bound: float | List[float] = 1, first_order_scale_lower_bound: float | List[float] = 1, power_lower_bound: float | List[float] = -1, power_upper_bound: float | List[float] = 1, pca_importance_lower_bound: float = 0, pca_importance_upper_bound: float = 10, config: Dict | None = None)[source]

This class provides methods to optimize the portfolio objective.

__init__(calculator: Calculator | LogarithmPCACalculator, direction: str | None = None, formula: str | None = None, warmup_formula: str | None = None, warmup_trials: int = 200, first_order: bool | None = False, power: bool | None = True, dirichlet: bool | None = False, weights_num: int | None = None, study_name: str | None = None, study_path: str | None = None, save_study: bool | None = True, checkpoint_path: str | None = None, first_order_with_scales: bool = True, first_order_lower_bound: float = 0.001, first_order_upper_bound: float = 1000000.0, free_style_lower_bound: float | List[float] = 0.001, free_style_upper_bound: float | List[float] = 1000000.0, base_weights: List[float] | None = None, base_weights_offset_ratio: float = 0.1, max_min_scale_ratio: float | None = None, first_order_scale_upper_bound: float | List[float] = 1, first_order_scale_lower_bound: float | List[float] = 1, power_lower_bound: float | List[float] = -1, power_upper_bound: float | List[float] = 1, pca_importance_lower_bound: float = 0, pca_importance_upper_bound: float = 10, config: Dict | None = None) → None[source]

Initialize with direction, weights_num, formula, and dirichlet.

Parameters:

first_order_with_scales (bool, optional) – Whether to use scales-control in the first-order objective. Defaults to True.
first_order_lower_bound (float, optional) – Lower bound for first order value. Defaults to 1e-3.
first_order_upper_bound (float, optional) – Upper bound for first order value. Defaults to 1e6.
power_lower_bound (Union[float, List[float]]) – Lower bound for power value. Defaults to -1.
power_upper_bound (Union[float, List[float]]) – Upper bound for power value. Defaults to 1.
pca_importance_lower_bound (float, optional) – Lower bound for pca importance value. Defaults to 0.
pca_importance_upper_bound (float, optional) – Upper bound for pca importance value. Defaults to 10.
first_order_scale_bound (Optional[float], optional) – Scale bound for first order value. Defaults to None.

Methods

`__init__`(calculator[, direction, formula, ...])	Initialize with direction, weights_num, formula, and dirichlet.
`add_evaluator`(flag, target_column[, ...])	Adds evaluators to the objective.
`build_logger`([process_id])	Constructs a logger for the optimization process.
`construct_weights`(trial)
`evaluate_custom_weights`(weights)	Evaluate the objective function with custom weights.
`evaluate_given_scores`(scores)	Evaluate the objective function with given scores.
`export_completed_formulas`([weights])	Exports the completed formulas by replacing weight placeholders in the formulas with actual values from the provided or default weights.
`get_weights_num`()	Returns the number of weights.
`objective`(trial)	Objective function to be optimized by optuna.
`optimize`(n_trials)	Optimizes the objective for a set number of trials.

Adds evaluators to the objective.

Parameters:

flag (str) – Type of calculator. Expected values include [“wuauc”, “portfolio”, “logmse”, …].
target_column (str) – The target column for calculation.
hyperparameter (Optional[float], optional) – Hyperparameter for the calculator. Defaults to None.
evaluator_property (Optional[str], optional) – Property of the evaluator. Defaults to None.
groupby (Optional[str], optional) – Grouping criteria. Defaults to None.

evaluate_custom_weights(weights: List[float]) → List[float][source]

Evaluate the objective function with custom weights.

Parameters:: weights (List[float]) – Custom weights to evaluate.

evaluate_given_scores(scores: List[float]) → List[float][source]

Evaluate the objective function with given scores.

Parameters:: scores (List[float]) – Scores to evaluate.

objective(trial: Trial) → float[source]

Objective function to be optimized by optuna.

Parameters:: trial (Trial) – Optuna trial instance.
Returns:: Computed objective value based on the provided trial.
Return type:: float

export_completed_formulas(weights: ndarray | None = None) → None[source]

Exports the completed formulas by replacing weight placeholders in the formulas with actual values from the provided or default weights.

If weights is not provided, it defaults to self.best_params. The method updates self.completed_formulas by replacing occurrences of weights[i] in the stored equations with corresponding values from the weight array.

Parameters:: weights (Optional[np.ndarray]) – An optional numpy array containing weight values to substitute in the formulas. If None, self.best_params is used.
Returns:: This method does not return a value but updates self.completed_formulas with the substituted equations.
Return type:: None