finol.optimization_layer.ParametersTuner

class finol.optimization_layer.ParametersTuner(load_dataset_output)

Class to tune model hyper-parameters using auto ML library.

參數:

load_dataset_output –

範例

>>> # Load dataset
>>> from finol.data_layer.dataset_loader import DatasetLoader
>>> load_dataset_output = DatasetLoader().load_dataset()
>>>
>>> # Tune model
>>> ParametersTuner(load_dataset_output=load_dataset_output).tune_parameters()

Methods

objective(trial)	Objective function for Optuna optimization.
sample_params(trial)	Samples model hyper-parameters for optimization.
select_pruner()	Selects and initializes an Optuna pruner based on the configuration.
select_sampler()	Selects and initializes an Optuna sampler based on the configuration.
tune_parameters()	Tune model hyper-parameters.

objective(trial)

Objective function for Optuna optimization. Trains the model using the sampled parameters and returns the validation loss.

參數:

trial (Trial) – Optuna Trial object.

回傳:

Validation loss.

回傳型別:

float

sample_params(trial)

Samples model hyper-parameters for optimization.

參數:

trial (Trial) – Optuna Trial object to sample the parameters.

回傳:

None

回傳型別:

None

select_pruner()

Selects and initializes an Optuna pruner based on the configuration.

回傳:

Initialized Optuna pruner object.

回傳型別:

object

select_sampler()

Selects and initializes an Optuna sampler based on the configuration.

回傳:

Initialized Optuna sampler object.

回傳型別:

object

tune_parameters()

Tune model hyper-parameters.

This method creates an Optuna study object, optimizes the objective function, visualizes the results, and prints out the optimization results.

回傳:

None

回傳型別:

None