streamline.modeling.basemodel module

class streamline.modeling.basemodel.BaseModel(model, model_name, cv_folds=3, scoring_metric='balanced_accuracy', metric_direction='maximize', random_state=None, cv=None, sampler=None, n_jobs=None)

Bases: object

Base Model Class for all ML Models

Parameters:

• model –

• model_name –

• cv_folds –

• scoring_metric –

• metric_direction –

• random_state –

• cv –

• sampler –

• n_jobs –

feature_importance()

Unimplemented feature importance function stub

fit(x_train, y_train, n_trails, timeout, feature_names=None)

Caller function to optimize

hyper_eval()

Hyper eval for objective function Returns: Returns hyper eval for objective function

model_evaluation(x_test, y_test)

Runs commands to gather all evaluations for later summaries and plots.

objective(trial, params=None)

Unimplemented objective function stub, needs to be overridden :param trial: optuna trial object :param params: dict of optional params or None

optimize(x_train, y_train, n_trails, timeout, feature_names=None)

Common model optimization function

Parameters:

• x_train – train data

• y_train – label data

• n_trails – number of optuna trials

• timeout – maximum time for optuna trial timeout

• feature_names – header/name of features

predict(x_in)

Function to predict with trained model :param x_in: input data

Returns: predictions y_pred