SparkMetric

class sparklightautoml.tasks.base.SparkMetric(name, metric_name, target_col=None, prediction_col=None, greater_is_better=True)

Bases: LAMLMetric

Spark version of metric function that implements function assessing prediction error.

__init__(name, metric_name, target_col=None, prediction_col=None, greater_is_better=True)

Parameters:

• name (str) – Name of metric.

• target_col (Optional[str]) – Name of column that stores targets

• prediction_col (Optional[str]) – Name of column that stores predictions

• greater_is_better (bool) – Whether or not higher metric value is better.