streamline.modeling.modeljob module

class streamline.modeling.modeljob.ModelJob(full_path, output_path, experiment_name, cv_count, class_label='Class', instance_label=None, scoring_metric='balanced_accuracy', metric_direction='maximize', n_trials=200, timeout=900, training_subsample=0, uniform_fi=False, save_plot=False, random_state=None)

Bases: Job

Parameters:

• full_path –

• output_path –

• experiment_name –

• cv_count –

• class_label –

• instance_label –

• scoring_metric –

• metric_direction –

• n_trials –

• timeout –

• uniform_fi –

• save_plot –

• random_state –

data_prep()

Loads target cv training dataset, separates class from features and removes instance labels.

static export_best_params(file_name, param_grid)

Exports the best hyperparameter scores to output file.

run(model)

Parameters:

model – model object

run_model(model)

Parameters:

model – model object

Returns: list of metrics [metric_list, fpr, tpr, roc_auc, prec, recall, prec_rec_auc, ave_prec, fi, probas]

save_runtime()

Save ML algorithm training and evaluation runtime for this phase.