mlrl.testbed.probability_calibration module

Author: Michael Rapp (michael.rapp.ml@gmail.com)

Provides classes for printing probability calibration models. The models can be written to one or several outputs, e.g., to the console or to a file.

class mlrl.testbed.probability_calibration.JointProbabilityCalibrationModelWriter(sinks: List[Sink])

Bases: ProbabilityCalibrationModelWriter

Allow to write textual representations of models for the calibration of joint probabilities to one or several sinks.

class CsvSink(output_dir: str, options: ~mlrl.common.options.Options = <mlrl.common.options.Options object>)

Bases: CsvSink

Allows to write textual representations of models for the calibration of joint probabilities to a CSV file.

class LogSink(options: ~mlrl.common.options.Options = <mlrl.common.options.Options object>)

Bases: LogSink

Allows to write textual representations of models for the calibration of joint probabilities to the console.

class mlrl.testbed.probability_calibration.MarginalProbabilityCalibrationModelWriter(sinks: List[Sink])

Bases: ProbabilityCalibrationModelWriter

Allow to write textual representations of models for the calibration of marginal probabilities to one or several sinks.

class CsvSink(output_dir: str, options: ~mlrl.common.options.Options = <mlrl.common.options.Options object>)

Bases: CsvSink

Allows to write textual representations of models for the calibration of marginal probabilities to a CSV file.

class LogSink(options: ~mlrl.common.options.Options = <mlrl.common.options.Options object>)

Bases: LogSink

Allows to write textual representations of models for the calibration of marginal probabilities to the console.

class mlrl.testbed.probability_calibration.ProbabilityCalibrationModelWriter(sinks: List[Sink], list_title: str)

Bases: OutputWriter, ABC

An abstract base class for all classes that allow to write textual representations of probability calibration models to one or several sinks.

class IsotonicProbabilityCalibrationModelFormattable(calibration_model: IsotonicProbabilityCalibrationModel, list_title: str)

Bases: IsotonicProbabilityCalibrationModelVisitor, Formattable, Tabularizable

Allows to create a textual representation of a model for the calibration of probabilities via isotonic regression.

format(options: Options, **_) → str

See mlrl.testbed.output_writer.Formattable.format()

tabularize(options: Options, **_) → List[Dict[str, str]] | None

See mlrl.testbed.output_writer.Tabularizable.tabularize()

visit_bin(list_index: int, threshold: float, probability: float)

See mlrl.common.cython.probability_calibration.IsotonicProbabilityCalibrationModelVisitor.visit_bin()

class NoProbabilityCalibrationModelFormattable

Bases: Formattable, Tabularizable

Allows to create a textual representation of a model for the calibration of probabilities that does not make any adjustments.

format(options: Options, **_) → str

See mlrl.testbed.output_writer.Formattable.format()

tabularize(options: Options, **_) → List[Dict[str, str]] | None

See mlrl.testbed.output_writer.Tabularizable.tabularize()