org.apache.spark.mllib.optimization

Class Gradient

java.lang.Object

org.apache.spark.mllib.optimization.Gradient

All Implemented Interfaces:

java.io.Serializable

Direct Known Subclasses:

HingeGradient, LeastSquaresGradient, LogisticGradient

public abstract class Gradient
extends java.lang.Object
implements scala.Serializable

:: DeveloperApi :: Class used to compute the gradient for a loss function, given a single data point.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
Gradient()

Method Summary

Methods

Modifier and Type	Method and Description
scala.Tuple2<Vector,java.lang.Object>	compute(Vector data, double label, Vector weights) Compute the gradient and loss given the features of a single data point.
abstract double	compute(Vector data, double label, Vector weights, Vector cumGradient) Compute the gradient and loss given the features of a single data point, add the gradient to a provided vector to avoid creating new objects, and return loss.

Methods inherited from class java.lang.Object

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

Constructor Detail

Gradient

public Gradient()

Method Detail

compute

public scala.Tuple2<Vector,java.lang.Object> compute(Vector data,
                                            double label,
                                            Vector weights)

Compute the gradient and loss given the features of a single data point.

Parameters:

data - features for one data point

label - label for this data point

weights - weights/coefficients corresponding to features

Returns:

(gradient: Vector, loss: Double)

compute

public abstract double compute(Vector data,
             double label,
             Vector weights,
             Vector cumGradient)

Compute the gradient and loss given the features of a single data point, add the gradient to a provided vector to avoid creating new objects, and return loss.

Parameters:

data - features for one data point

label - label for this data point

weights - weights/coefficients corresponding to features

cumGradient - the computed gradient will be added to this vector

Returns:

loss