Module datarush.analytics

Package com.pervasive.datarush.analytics.pmml

Class PMMLModelSpec

java.lang.Object

com.pervasive.datarush.analytics.pmml.PMMLModelSpec

All Implemented Interfaces:

Serializable

public final class PMMLModelSpec
extends Object
implements Serializable

Meta-information associated with a given PMMLModel.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor	Description
PMMLModelSpec(RecordTokenType inputType, List<String> learningCols, List<String> targetCols)	Creates a PMMLModelSpec for the given input

Method Summary

Modifier and Type	Method	Description
static PMMLModelSpec	forAll(RecordTokenType inputType, String... targetColumns)	Convenience method to create a model spec where the learning columns are all input columns minus the target columns
static PMMLModelSpec	forSelected(RecordTokenType inputType, List<String> selectedColumns, String... targetColumns)	Convenience method to create a model spec where the learning columns are the specified selected columns minus the set of target columns.
RecordTokenType	getInputType()	Returns the input type for which we are constructing a model
List<String>	getLearningCols()	Returns the learning columns (i.e.
List<String>	getTargetCols()	Returns the target columns (i.e.
RecordTokenType	getTargetType()	Returns the target type.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

PMMLModelSpec

public PMMLModelSpec(RecordTokenType inputType,
                     List<String> learningCols,
                     List<String> targetCols)

Creates a PMMLModelSpec for the given input

Parameters:

inputType - the input type being learned

learningCols - the learning columns (i.e. independent variables)

targetCols - the target columns (i.e. dependent variables)

Method Detail

forAll

public static PMMLModelSpec forAll(RecordTokenType inputType,
                                   String... targetColumns)

Convenience method to create a model spec where the learning columns are all input columns minus the target columns

Parameters:

inputType - the full input type

targetColumns - the target columns

Returns:

a new model spec

forSelected

public static PMMLModelSpec forSelected(RecordTokenType inputType,
                                        List<String> selectedColumns,
                                        String... targetColumns)

Convenience method to create a model spec where the learning columns are the specified selected columns minus the set of target columns.

Parameters:

inputType - the input type

selectedColumns - the selected columns. If null or empty the learning columns are all columns minus the target columns. If non-empty the learning columns are selected columns minus the target columns.

targetColumns - the target columns

Returns:

a PMMLModelSpec

getInputType

public RecordTokenType getInputType()

Returns the input type for which we are constructing a model

Returns:

the input type for which we are constructing a model

getLearningCols

public List<String> getLearningCols()

Returns the learning columns (i.e. independent variables)

Returns:

the learning columns

getTargetCols

public List<String> getTargetCols()

Returns the target columns (i.e. dependent variables)

Returns:

the target columns

getTargetType

public RecordTokenType getTargetType()

Returns the target type. This is the target columns selected from the input type.

Returns:

the target type