com.pervasive.datarush.analytics.util

Class AbstractPredictor

java.lang.Object

com.pervasive.datarush.operators.AbstractLogicalOperator

com.pervasive.datarush.operators.StreamingOperator

com.pervasive.datarush.operators.ExecutableOperator

com.pervasive.datarush.analytics.util.AbstractPredictor

All Implemented Interfaces:

LogicalOperator

Direct Known Subclasses:

ClusterPredictor, DecisionTreePredictor, LogisticRegressionPredictor, NaiveBayesPredictor, RegressionPredictor, SVMPredictor

public abstract class AbstractPredictor
extends ExecutableOperator

Common base class for "predictor" processes. Predictors all take a PMML input, a record input and produce a record output. In addition, all predictors are parallel.

Constructor Summary

Constructors

Constructor and Description
AbstractPredictor()

Method Summary

Methods

Modifier and Type	Method and Description
protected void	computeMetadata(StreamingMetadataContext ctx) Default implementation of computeMetadata.
protected void	execute(ExecutionContext ctx) Invokes execute(PMMLModel, RecordValued, ScalarSettable[]) and performs an output.pushEndOfData()
protected abstract void	execute(PMMLModel model, RecordValued input, ScalarSettable[] predictedFields) Called to perform prediction.
RecordPort	getInput() The input data.
PMMLPort	getModel() The input for supplying the PMML model
RecordPort	getOutput() The original data with prediction appended
protected abstract RecordTokenType	predictedType(PMMLModelSpec modelSpec) Given the model spec, returns the predicted type.
protected void	pushPrediction() Pushes a prediction row consisting of input fields plus predicted fields
protected boolean	stepNext() Steps to the next input row, returning false at end of data

Methods inherited from class com.pervasive.datarush.operators.ExecutableOperator

cloneForExecution, getNumInputCopies, getPortSettings, handleInactiveOutput

Methods inherited from class com.pervasive.datarush.operators.AbstractLogicalOperator

disableParallelism, getInputPorts, getOutputPorts, newInput, newInput, newOutput, newRecordInput, newRecordInput, newRecordOutput, notifyError

Methods inherited from class java.lang.Object

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

Constructor Detail

AbstractPredictor

public AbstractPredictor()

Method Detail

getInput

public final RecordPort getInput()

The input data. Any fields specified in the PMML model must be present in the input data. Additional fields will be ignored.

Returns:

the input data

getModel

public final PMMLPort getModel()

The input for supplying the PMML model

Returns:

the input for the pmml model

getOutput

public RecordPort getOutput()

The original data with prediction appended

Returns:

the original data with the prediction appended

computeMetadata

protected void computeMetadata(StreamingMetadataContext ctx)

Default implementation of computeMetadata.

1. Output type is set to input type plus predictedType
2. Input data ordering ( if ordered ) is preserved
3. Input data partitioning ( if partitioned ) is preserved

Specified by:

computeMetadata in class StreamingOperator

Parameters:

ctx - the context

stepNext

protected final boolean stepNext()

Steps to the next input row, returning false at end of data

Returns:

true if there is still more data

pushPrediction

protected final void pushPrediction()

Pushes a prediction row consisting of input fields plus predicted fields

execute

protected final void execute(ExecutionContext ctx)

Invokes execute(PMMLModel, RecordValued, ScalarSettable[]) and performs an output.pushEndOfData()

Specified by:

execute in class ExecutableOperator

Parameters:

ctx - context in which to lookup physical ports bound to logical ports

predictedType

protected abstract RecordTokenType predictedType(PMMLModelSpec modelSpec)

Given the model spec, returns the predicted type. This should not include the input type ( the input is automatically prepended to the type that is returned )

Parameters:

modelSpec - the model metadata

Returns:

the predicted type

execute

protected abstract void execute(PMMLModel model,
           RecordValued input,
           ScalarSettable[] predictedFields)

Called to perform prediction. Subclasses are expected to loop over the input by calling stepNext(). For each row of input, subclasses should first set the predicted values in the predictedFields array and then invoke pushPrediction(). Subclasses should not invoke pushEndOfData since that is automatically handled by the base class.

Parameters:

model - The input PMML model

input - The input data

predictedFields - An array of fields that reference the predicted field locations. The array positionally corresponds to the type returned by predictedType(PMMLModelSpec).