org.apache.spark.mllib.classification

Class StreamingLogisticRegressionWithSGD

java.lang.Object

org.apache.spark.mllib.regression.StreamingLinearAlgorithm<LogisticRegressionModel,LogisticRegressionWithSGD>

org.apache.spark.mllib.classification.StreamingLogisticRegressionWithSGD

All Implemented Interfaces:

java.io.Serializable

public class StreamingLogisticRegressionWithSGD
extends StreamingLinearAlgorithm<LogisticRegressionModel,LogisticRegressionWithSGD>
implements scala.Serializable

Train or predict a logistic regression model on streaming data. Training uses Stochastic Gradient Descent to update the model based on each new batch of incoming data from a DStream (see LogisticRegressionWithSGD for model equation)

Each batch of data is assumed to be an RDD of LabeledPoints. The number of data points per batch can vary, but the number of features must be constant. An initial weight vector must be provided.

Use a builder pattern to construct a streaming logistic regression analysis in an application, like:

  val model = new StreamingLogisticRegressionWithSGD()
    .setStepSize(0.5)
    .setNumIterations(10)
    .setInitialWeights(Vectors.dense(...))
    .trainOn(DStream)
 

Constructor Summary

Constructors

Constructor and Description
StreamingLogisticRegressionWithSGD() Construct a StreamingLogisticRegression object with default parameters: {stepSize: 0.1, numIterations: 50, miniBatchFraction: 1.0, regParam: 0.0}.

Method Summary

Methods

Modifier and Type	Method and Description
protected LogisticRegressionWithSGD	algorithm() The algorithm to use for updating.
protected scala.Option<LogisticRegressionModel>	model() The model to be updated and used for prediction.
StreamingLogisticRegressionWithSGD	setInitialWeights(Vector initialWeights) Set the initial weights.
StreamingLogisticRegressionWithSGD	setMiniBatchFraction(double miniBatchFraction) Set the fraction of each batch to use for updates.
StreamingLogisticRegressionWithSGD	setNumIterations(int numIterations) Set the number of iterations of gradient descent to run per update.
StreamingLogisticRegressionWithSGD	setRegParam(double regParam) Set the regularization parameter.
StreamingLogisticRegressionWithSGD	setStepSize(double stepSize) Set the step size for gradient descent.

Methods inherited from class org.apache.spark.mllib.regression.StreamingLinearAlgorithm

latestModel, predictOn, predictOn, predictOnValues, predictOnValues, trainOn, trainOn

Methods inherited from class java.lang.Object

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

Constructor Detail

StreamingLogisticRegressionWithSGD

public StreamingLogisticRegressionWithSGD()

Construct a StreamingLogisticRegression object with default parameters: {stepSize: 0.1, numIterations: 50, miniBatchFraction: 1.0, regParam: 0.0}. Initial weights must be set before using trainOn or predictOn (see StreamingLinearAlgorithm)

Method Detail

algorithm

protected LogisticRegressionWithSGD algorithm()

Description copied from class: StreamingLinearAlgorithm

The algorithm to use for updating.

Specified by:

algorithm in class StreamingLinearAlgorithm<LogisticRegressionModel,LogisticRegressionWithSGD>

model

protected scala.Option<LogisticRegressionModel> model()

Description copied from class: StreamingLinearAlgorithm

The model to be updated and used for prediction.

Specified by:

model in class StreamingLinearAlgorithm<LogisticRegressionModel,LogisticRegressionWithSGD>

setStepSize

public StreamingLogisticRegressionWithSGD setStepSize(double stepSize)

Set the step size for gradient descent. Default: 0.1.

setNumIterations

public StreamingLogisticRegressionWithSGD setNumIterations(int numIterations)

Set the number of iterations of gradient descent to run per update. Default: 50.

setMiniBatchFraction

public StreamingLogisticRegressionWithSGD setMiniBatchFraction(double miniBatchFraction)

Set the fraction of each batch to use for updates. Default: 1.0.

setRegParam

public StreamingLogisticRegressionWithSGD setRegParam(double regParam)

Set the regularization parameter. Default: 0.0.

setInitialWeights

public StreamingLogisticRegressionWithSGD setInitialWeights(Vector initialWeights)

Set the initial weights. Default: [0.0, 0.0].