Class SimplePerceptronSequenceTrainer<T>

Namespace

NOpenNLP.Tools.Ml.Perceptron

Assembly

NOpenNLP.Tools.dll

Trains models for sequences using the perceptron algorithm. Each outcome is represented as a binary perceptron classifier. This supports standard (integer) weighting as well as average weighting. Sequence information is used in a simplified way to that described in: Discriminative Training Methods for Hidden Markov Models: Theory and Experiments with the Perceptron Algorithm. Michael Collins, EMNLP 2002. Specifically only updates are applied to tokens which were incorrectly tagged by a sequence tagger rather than to all features across the sequence which differ from the training sequence.

public class SimplePerceptronSequenceTrainer<T> : AbstractEventModelSequenceTrainer<T>, IEventModelSequenceTrainer<T>

Type Parameters

T

The type of the object which is the source of each sequence.

Inheritance

object

AbstractTrainer

AbstractEventModelSequenceTrainer<T>

SimplePerceptronSequenceTrainer<T>

Implements

IEventModelSequenceTrainer<T>

Inherited Members

AbstractEventModelSequenceTrainer<T>.Train(ISequenceStream<T>)

AbstractTrainer.ALGORITHM_PARAM

AbstractTrainer.TRAINER_TYPE_PARAM

AbstractTrainer.CUTOFF_PARAM

AbstractTrainer.CUTOFF_DEFAULT

AbstractTrainer.ITERATIONS_PARAM

AbstractTrainer.ITERATIONS_DEFAULT

AbstractTrainer.VERBOSE_PARAM

AbstractTrainer.VERBOSE_DEFAULT

AbstractTrainer.trainingParameters

AbstractTrainer.reportMap

AbstractTrainer.printMessages

AbstractTrainer.Init(TrainingParameters, IDictionary<string, string>)

AbstractTrainer.Algorithm

AbstractTrainer.Cutoff

AbstractTrainer.Iterations

AbstractTrainer.AddToReport(string, string)

AbstractTrainer.Display(string)

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Constructors

SimplePerceptronSequenceTrainer()

public SimplePerceptronSequenceTrainer()

Fields

PERCEPTRON_SEQUENCE_VALUE

public const string PERCEPTRON_SEQUENCE_VALUE = "PERCEPTRON_SEQUENCE"

Field Value

string

Methods

DoTrain(ISequenceStream<T>)

Trains a model from the given sequence stream.

public override IMaxentModel DoTrain(ISequenceStream<T> events)

Parameters

events ISequenceStream<T>

Returns

IMaxentModel

NextIteration(int)

public virtual void NextIteration(int iteration)

Parameters

iteration int

TrainModel(int, ISequenceStream<T>, int, bool)

public virtual AbstractModel TrainModel(int iterations, ISequenceStream<T> sequenceStream, int cutoff, bool useAverage)

Parameters

iterations int

sequenceStream ISequenceStream<T>

cutoff int

useAverage bool

Returns

AbstractModel

Validate()

Checks the parameters. If a subclass overrides this, it should call the base implementation.

public override void Validate()

Exceptions

ArgumentException

Thrown if a parameter is not valid.