smile.sequence

Class CRF.Trainer

java.lang.Object

smile.sequence.CRF.Trainer

Enclosing class:

CRF

public static class CRF.Trainer
extends java.lang.Object

Trainer for CRF.

Constructor Summary

Constructors

Constructor and Description
Trainer(smile.data.Attribute[] attributes, int numClasses) Constructor.
Trainer(int numFeatures, int numClasses) Constructor.

Method Summary

Modifier and Type	Method and Description
CRF.Trainer	setLearningRate(double eta)
CRF.Trainer	setMaxNodes(int maxLeaves) Sets the maximum number of leaf nodes in the tree.
CRF.Trainer	setNumTrees(int iters)
CRF	train(double[][][] sequences, int[][] labels)
CRF	train(int[][][] sequences, int[][] labels)

Methods inherited from class java.lang.Object

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

Constructor Detail

Trainer

public Trainer(smile.data.Attribute[] attributes,
               int numClasses)

Constructor.

Parameters:

numClasses - the maximum number of classes.

Trainer

public Trainer(int numFeatures,
               int numClasses)

Constructor.

Parameters:

numFeatures - the number of sparse binary features.

numClasses - the maximum number of classes.

Method Detail

setMaxNodes

public CRF.Trainer setMaxNodes(int maxLeaves)

Sets the maximum number of leaf nodes in the tree.

Parameters:

maxLeaves - the maximum number of leaf nodes in the tree.

setLearningRate

public CRF.Trainer setLearningRate(double eta)

setNumTrees

public CRF.Trainer setNumTrees(int iters)

train

public CRF train(double[][][] sequences,
                 int[][] labels)

train

public CRF train(int[][][] sequences,
                 int[][] labels)