lookout.style.format.optimizer

Optimize base model hyper-parameters.

Module Contents

class lookout.style.format.optimizer.Optimizer(cv:int, n_iter:int, n_jobs:Optional[int], random_state:int, base_model_name_categories:Sequence[str], max_depth_categories:Sequence[Optional[int]], max_features_categories:Sequence[Optional[str]], min_samples_leaf_min:int, min_samples_leaf_max:int, min_samples_split_min:int, min_samples_split_max:int)

Optimize base model hyper-parameters.

_log

optimize(self, X:csr_matrix, y:numpy.ndarray)

Conduct hyper-parameters search to find the best base model given the data.

Parameters:

• X – Sparse feature matrix.
• y – Labels numpy array.

Returns:

Best base model score and parameters.

_cost(self, *, X:csr_matrix, y:numpy.ndarray, **params)

class lookout.style.format.optimizer._VerboseLogCallback(log)

Callback to control the verbosity and log output properly.

Adopted from skopt library, VerboseCallback class.

__call__(self, res)

Call callback method.

Parameters:res – The optimization as a OptimizeResult object. Returns:None