cvopt.utils.extract_params¶
-
extract_params
(logdir, model_id, target_index, feature_groups=None)[source]¶ Extract parameters from cvopt logfile.
Parameters: - logdir (str.) – cvopt’s log directory path.
- model_id (str.) – cvopt’s model id.
- target_index (int.) – Logfile index(start:0). Parameters correspond to index is extracted.
- feature_groups (array-like, shape = (n_samples,) or None, default=None.) –
cvopt feature_groups. When feature_groups is None, feature_param and feature_select_flag in returns is None.
feature select flag is bool vector. If this value is True, optimizer recommend using corresponding column.
Returns: - estimator_params (dict) – estimator parameters of the target model.
- feature_params (dict or None) – feature parameters of the target model.
- feature_select_flag (numpy array or None) – feature select flag of the target model.