Examples of opennlp.tools.ml.EventTrainer

opennlp.tools.ml.EventTrainer


    if (TrainerType.EVENT_MODEL_TRAINER.equals(trainerType)) {
      ObjectStream<Event> eventStream = new NameFinderEventStream(samples, type,
              factory.createContextGenerator(), factory.createSequenceCodec());


      EventTrainer trainer = TrainerFactory.getEventTrainer(trainParams.getSettings(), manifestInfoEntries);
      nameFinderModel = trainer.train(eventStream);
    } // TODO: Maybe it is not a good idea, that these two don't use the context generator ?!
    // These also don't use the sequence codec ?!
    else if (TrainerType.EVENT_MODEL_SEQUENCE_TRAINER.equals(trainerType)) {
      NameSampleSequenceStream ss = new NameSampleSequenceStream(samples, factory.createContextGenerator());


      EventModelSequenceTrainer trainer = TrainerFactory.getEventModelSequenceTrainer(
              trainParams.getSettings(), manifestInfoEntries);
      nameFinderModel = trainer.train(ss);
    } else if (TrainerType.SEQUENCE_TRAINER.equals(trainerType)) {
      SequenceTrainer trainer = TrainerFactory.getSequenceModelTrainer(
              trainParams.getSettings(), manifestInfoEntries);


      NameSampleSequenceStream ss = new NameSampleSequenceStream(samples, factory.createContextGenerator(), false);
      seqModel = trainer.train(ss);
    } else {
      throw new IllegalStateException("Unexpected trainer type!");
    }


    if (seqModel != null) {

View Full Code Here


    if (TrainerType.EVENT_MODEL_TRAINER.equals(trainerType)) {
      ObjectStream<Event> eventStream = new NameFinderEventStream(samples, type,
              new DefaultNameContextGenerator(featureGenerator), new BioCodec());


      EventTrainer trainer = TrainerFactory.getEventTrainer(trainParams.getSettings(), manifestInfoEntries);
      nameFinderModel = trainer.train(eventStream);
    } else if (TrainerType.EVENT_MODEL_SEQUENCE_TRAINER.equals(trainerType)) {
      NameSampleSequenceStream ss = new NameSampleSequenceStream(samples, featureGenerator);


      EventModelSequenceTrainer trainer = TrainerFactory.getEventModelSequenceTrainer(
              trainParams.getSettings(), manifestInfoEntries);
      nameFinderModel = trainer.train(ss);
    } else if (TrainerType.SEQUENCE_TRAINER.equals(trainerType)) {
      SequenceTrainer trainer = TrainerFactory.getSequenceModelTrainer(
              trainParams.getSettings(), manifestInfoEntries);


      NameSampleSequenceStream ss = new NameSampleSequenceStream(samples, featureGenerator, false);
      seqModel = trainer.train(ss);
    } else {
      throw new IllegalStateException("Unexpected trainer type!");
    }


     // TODO: Pass the sequence codec down to the model! We will just store the class

View Full Code Here

    MaxentModel posModel = null;
    SequenceClassificationModel<String> seqPosModel = null;
    if (TrainerType.EVENT_MODEL_TRAINER.equals(trainerType)) {
      ObjectStream<Event> es = new POSSampleEventStream(samples, contextGenerator);


      EventTrainer trainer = TrainerFactory.getEventTrainer(trainParams.getSettings(),
          manifestInfoEntries);
      posModel = trainer.train(es);
    }
    else if (TrainerType.EVENT_MODEL_SEQUENCE_TRAINER.equals(trainerType)) {
      POSSampleSequenceStream ss = new POSSampleSequenceStream(samples, contextGenerator);
      EventModelSequenceTrainer trainer = TrainerFactory.getEventModelSequenceTrainer(trainParams.getSettings(),
          manifestInfoEntries);
      posModel = trainer.train(ss);
    }
    else if (TrainerType.SEQUENCE_TRAINER.equals(trainerType)) {
      SequenceTrainer trainer = TrainerFactory.getSequenceModelTrainer(
          trainParams.getSettings(), manifestInfoEntries);


      // TODO: This will probably cause issue, since the feature generator uses the outcomes array


      POSSampleSequenceStream ss = new POSSampleSequenceStream(samples, contextGenerator);
      seqPosModel = trainer.train(ss);
    }
    else {
      throw new IllegalArgumentException("Trainer type is not supported: " + trainerType);
    }

View Full Code Here

    MaxentModel chunkerModel = null;
    SequenceClassificationModel<String> seqChunkerModel = null;


    if (TrainerType.EVENT_MODEL_TRAINER.equals(trainerType)) {
      ObjectStream<Event> es = new ChunkerEventStream(in, factory.getContextGenerator());
      EventTrainer trainer = TrainerFactory.getEventTrainer(mlParams.getSettings(),
          manifestInfoEntries);
      chunkerModel = trainer.train(es);
    }
    else if (TrainerType.SEQUENCE_TRAINER.equals(trainerType)) {
      SequenceTrainer trainer = TrainerFactory.getSequenceModelTrainer(
          mlParams.getSettings(), manifestInfoEntries);


      // TODO: This will probably cause issue, since the feature generator uses the outcomes array


      ChunkSampleSequenceStream ss = new ChunkSampleSequenceStream(in, factory.getContextGenerator());
      seqChunkerModel = trainer.train(ss);
    }
    else {
      throw new IllegalArgumentException("Trainer type is not supported: " + trainerType);
    }

View Full Code Here

      throws IOException {


    if(!TrainerFactory.isSupportEvent(trainParams)) {
      throw new IllegalArgumentException("EventTrain is not supported");
    }
    EventTrainer trainer = TrainerFactory.getEventTrainer(trainParams, reportMap);


    return trainer.train(events);
  }

View Full Code Here

TOP

Related Classes of opennlp.tools.ml.EventTrainer

opennlp.tools.chunker.ChunkerME

opennlp.tools.ml.model.TrainUtil

opennlp.tools.namefind.NameFinderME

opennlp.tools.postag.POSTaggerME

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.