Basic usage.md

Basic usage

As part of our persona-driven development BlueCast offers two ways of configuration:

• a non-programmatic way which uses widgets for most important options
• a programmatic way with ull control and flexibility

• Basic usage
  • Non-programmatic way for starters
  • Programmatic way

Non-programmatic way for starters

The non-programmatic way should be helpful for new BlueCast users or ml practitioners with little ml experience. The interface can be used like this:

from bluecast.blueprints.welcome import WelcomeToBlueCast

welcome = WelcomeToBlueCast()
welcome.automl_configurator()

# here users can chose from the given options and click 'submit' to create the instance
# after submit the automl instance can be retrieved and used like:
automl = welcome.automl_instance
automl.fit(df_train, target_col="target")
y_hat = automl.predict(df_val)

Programmatic way

The programmatic way is suitable for experienced users.

The module blueprints contains the main functionality of the library. The main entry point is the Blueprint class. It already includes needed preprocessing (including some convenience functionality like feature type detection) and model hyperparameter tuning.

from bluecast.blueprints.cast import BlueCast

automl = BlueCast(
        class_problem="binary",
    )

automl.fit(df_train, target_col="target")
y_probs, y_classes = automl.predict(df_val)

# from version 0.95 also predict_proba is directly available (also for BlueCastCV)
y_probs = automl.predict_proba(df_val)

BlueCast has simple utilities to save and load your pipeline:

from bluecast.general_utils.general_utils import save_to_production, load_for_production

# save pipeline including tracker
save_to_production(automl, "/kaggle/working/", "bluecast_cv_pipeline")

# in production or for further experiments this can be loaded again
automl = load_for_production("/kaggle/working/", "bluecast_cv_pipeline")

Since version 0.80 BlueCast offers regression as well:

from bluecast.blueprints.cast_regression import BlueCastRegression

automl = BlueCast(
        class_problem="regression",
    )

automl.fit(df_train, target_col="target")
y_hat = automl.predict(df_val)

During classification taks the target labels might arrive as strings. BlueCast will return these as numeric representations by default at the monment. However the predict method accepts a parameter return_original_labels to return the original labels.

from bluecast.blueprints.cast_regression import BlueCastRegression

automl = BlueCast(
        class_problem="binary",
    )

automl.fit(df_train, target_col="target")
y_probs, y_classes = automl.predict(df_val, return_original_labels=True)

The default behaviour might change in a future release.