synapse.ml.exploratory package

Submodules

synapse.ml.exploratory.AggregateBalanceMeasure module

class synapse.ml.exploratory.AggregateBalanceMeasure.AggregateBalanceMeasure(java_obj=None, epsilon=1.0, errorTolerance=1e-12, outputCol='AggregateBalanceMeasure', sensitiveCols=None, verbose=False)[source]

Bases: ComplexParamsMixin, JavaMLReadable, JavaMLWritable, JavaTransformer

Parameters:

epsilon¶ (float) – Epsilon value for Atkinson Index. Inverse of alpha (1 - alpha).
errorTolerance¶ (float) – Error tolerance value for Atkinson Index.
outputCol¶ (str) – output column name
sensitiveCols¶ (list) – Sensitive columns to use.
verbose¶ (bool) – Whether to show intermediate measures and calculations, such as Positive Rate.

epsilon = Param(parent='undefined', name='epsilon', doc='Epsilon value for Atkinson Index. Inverse of alpha (1 - alpha).')

errorTolerance = Param(parent='undefined', name='errorTolerance', doc='Error tolerance value for Atkinson Index.')

getEpsilon()[source]

Returns:: Epsilon value for Atkinson Index. Inverse of alpha (1 - alpha).
Return type:: epsilon

getErrorTolerance()[source]

Returns:: Error tolerance value for Atkinson Index.
Return type:: errorTolerance

static getJavaPackage()[source]: Returns package name String.

getOutputCol()[source]

Returns:: output column name
Return type:: outputCol

getSensitiveCols()[source]

Returns:: Sensitive columns to use.
Return type:: sensitiveCols

getVerbose()[source]

Returns:: Whether to show intermediate measures and calculations, such as Positive Rate.
Return type:: verbose

outputCol = Param(parent='undefined', name='outputCol', doc='output column name')

classmethod read()[source]: Returns an MLReader instance for this class.

sensitiveCols = Param(parent='undefined', name='sensitiveCols', doc='Sensitive columns to use.')

setEpsilon(value)[source]

Parameters:: epsilon¶ – Epsilon value for Atkinson Index. Inverse of alpha (1 - alpha).

setErrorTolerance(value)[source]

Parameters:: errorTolerance¶ – Error tolerance value for Atkinson Index.

setOutputCol(value)[source]

Parameters:: outputCol¶ – output column name

setParams(epsilon=1.0, errorTolerance=1e-12, outputCol='AggregateBalanceMeasure', sensitiveCols=None, verbose=False)[source]: Set the (keyword only) parameters

setSensitiveCols(value)[source]

Parameters:: sensitiveCols¶ – Sensitive columns to use.

setVerbose(value)[source]

Parameters:: verbose¶ – Whether to show intermediate measures and calculations, such as Positive Rate.

verbose = Param(parent='undefined', name='verbose', doc='Whether to show intermediate measures and calculations, such as Positive Rate.')

synapse.ml.exploratory.DistributionBalanceMeasure module

class synapse.ml.exploratory.DistributionBalanceMeasure.DistributionBalanceMeasure(java_obj=None, featureNameCol='FeatureName', outputCol='DistributionBalanceMeasure', referenceDistribution=None, sensitiveCols=None, verbose=False)[source]

Bases: ComplexParamsMixin, JavaMLReadable, JavaMLWritable, JavaTransformer

Parameters:

featureNameCol¶ (str) – Output column name for feature names.
outputCol¶ (str) – output column name
referenceDistribution¶ (object) – An ordered list of reference distributions that correspond to each of the sensitive columns.
sensitiveCols¶ (list) – Sensitive columns to use.
verbose¶ (bool) – Whether to show intermediate measures and calculations, such as Positive Rate.

featureNameCol = Param(parent='undefined', name='featureNameCol', doc='Output column name for feature names.')

getFeatureNameCol()[source]

Returns:: Output column name for feature names.
Return type:: featureNameCol

static getJavaPackage()[source]: Returns package name String.

getOutputCol()[source]

Returns:: output column name
Return type:: outputCol

getReferenceDistribution()[source]

Returns:: An ordered list of reference distributions that correspond to each of the sensitive columns.
Return type:: referenceDistribution

getSensitiveCols()[source]

Returns:: Sensitive columns to use.
Return type:: sensitiveCols

getVerbose()[source]

Returns:: Whether to show intermediate measures and calculations, such as Positive Rate.
Return type:: verbose

outputCol = Param(parent='undefined', name='outputCol', doc='output column name')

classmethod read()[source]: Returns an MLReader instance for this class.

referenceDistribution = Param(parent='undefined', name='referenceDistribution', doc='An ordered list of reference distributions that correspond to each of the sensitive columns.')

sensitiveCols = Param(parent='undefined', name='sensitiveCols', doc='Sensitive columns to use.')

setFeatureNameCol(value)[source]

Parameters:: featureNameCol¶ – Output column name for feature names.

setOutputCol(value)[source]

Parameters:: outputCol¶ – output column name

setParams(featureNameCol='FeatureName', outputCol='DistributionBalanceMeasure', referenceDistribution=None, sensitiveCols=None, verbose=False)[source]: Set the (keyword only) parameters

setReferenceDistribution(value)[source]

Parameters:: referenceDistribution¶ – An ordered list of reference distributions that correspond to each of the sensitive columns.

setSensitiveCols(value)[source]

Parameters:: sensitiveCols¶ – Sensitive columns to use.

setVerbose(value)[source]

Parameters:: verbose¶ – Whether to show intermediate measures and calculations, such as Positive Rate.

verbose = Param(parent='undefined', name='verbose', doc='Whether to show intermediate measures and calculations, such as Positive Rate.')

synapse.ml.exploratory.FeatureBalanceMeasure module

class synapse.ml.exploratory.FeatureBalanceMeasure.FeatureBalanceMeasure(java_obj=None, classACol='ClassA', classBCol='ClassB', featureNameCol='FeatureName', labelCol='label', outputCol='FeatureBalanceMeasure', sensitiveCols=None, verbose=False)[source]

Bases: ComplexParamsMixin, JavaMLReadable, JavaMLWritable, JavaTransformer

Parameters:

classACol¶ (str) – Output column name for the first feature value to compare.
classBCol¶ (str) – Output column name for the second feature value to compare.
featureNameCol¶ (str) – Output column name for feature names.
labelCol¶ (str) – label column name
outputCol¶ (str) – output column name
sensitiveCols¶ (list) – Sensitive columns to use.
verbose¶ (bool) – Whether to show intermediate measures and calculations, such as Positive Rate.

classACol = Param(parent='undefined', name='classACol', doc='Output column name for the first feature value to compare.')

classBCol = Param(parent='undefined', name='classBCol', doc='Output column name for the second feature value to compare.')

featureNameCol = Param(parent='undefined', name='featureNameCol', doc='Output column name for feature names.')

getClassACol()[source]

Returns:: Output column name for the first feature value to compare.
Return type:: classACol

getClassBCol()[source]

Returns:: Output column name for the second feature value to compare.
Return type:: classBCol

getFeatureNameCol()[source]

Returns:: Output column name for feature names.
Return type:: featureNameCol

static getJavaPackage()[source]: Returns package name String.

getLabelCol()[source]

Returns:: label column name
Return type:: labelCol

getOutputCol()[source]

Returns:: output column name
Return type:: outputCol

getSensitiveCols()[source]

Returns:: Sensitive columns to use.
Return type:: sensitiveCols

getVerbose()[source]

Returns:: Whether to show intermediate measures and calculations, such as Positive Rate.
Return type:: verbose

labelCol = Param(parent='undefined', name='labelCol', doc='label column name')

outputCol = Param(parent='undefined', name='outputCol', doc='output column name')

classmethod read()[source]: Returns an MLReader instance for this class.

sensitiveCols = Param(parent='undefined', name='sensitiveCols', doc='Sensitive columns to use.')

setClassACol(value)[source]

Parameters:: classACol¶ – Output column name for the first feature value to compare.

setClassBCol(value)[source]

Parameters:: classBCol¶ – Output column name for the second feature value to compare.

setFeatureNameCol(value)[source]

Parameters:: featureNameCol¶ – Output column name for feature names.

setLabelCol(value)[source]

Parameters:: labelCol¶ – label column name

setOutputCol(value)[source]

Parameters:: outputCol¶ – output column name

setParams(classACol='ClassA', classBCol='ClassB', featureNameCol='FeatureName', labelCol='label', outputCol='FeatureBalanceMeasure', sensitiveCols=None, verbose=False)[source]: Set the (keyword only) parameters

setSensitiveCols(value)[source]

Parameters:: sensitiveCols¶ – Sensitive columns to use.

setVerbose(value)[source]

Parameters:: verbose¶ – Whether to show intermediate measures and calculations, such as Positive Rate.

verbose = Param(parent='undefined', name='verbose', doc='Whether to show intermediate measures and calculations, such as Positive Rate.')

Module contents

SynapseML is an ecosystem of tools aimed towards expanding the distributed computing framework Apache Spark in several new directions. SynapseML adds many deep learning and data science tools to the Spark ecosystem, including seamless integration of Spark Machine Learning pipelines with Microsoft Cognitive Toolkit (CNTK), LightGBM and OpenCV. These tools enable powerful and highly-scalable predictive and analytical models for a variety of datasources.

SynapseML also brings new networking capabilities to the Spark Ecosystem. With the HTTP on Spark project, users can embed any web service into their SparkML models. In this vein, SynapseML provides easy to use SparkML transformers for a wide variety of Microsoft Cognitive Services. For production grade deployment, the Spark Serving project enables high throughput, sub-millisecond latency web services, backed by your Spark cluster.

SynapseML requires Scala 2.12, Spark 3.0+, and Python 3.6+.