How to use

opennlp.tools.lemmatizer.LemmatizerEvaluator

constructor

LemmatizerEvaluator evaluator = new LemmatizerEvaluator(
  new opennlp.tools.lemmatizer.LemmatizerME(model),
  missclassifiedListener, reportListener);

/**
 * Checks the evaluator results against the results got using the conlleval,
 * available at http://www.cnts.ua.ac.be/conll2000/chunking/output.html but
 * containing lemmas instead of chunks.
 *
 * @throws IOException
 */
@Test
public void testEvaluator() throws IOException {
 String inPredicted = "opennlp/tools/lemmatizer/output.txt";
 String inExpected = "opennlp/tools/lemmatizer/output.txt";
 String encoding = "UTF-8";
 DummyLemmaSampleStream predictedSample = new DummyLemmaSampleStream(
   new PlainTextByLineStream(
    new MockInputStreamFactory(new File(inPredicted)), encoding), true);
 DummyLemmaSampleStream expectedSample = new DummyLemmaSampleStream(
   new PlainTextByLineStream(
    new MockInputStreamFactory(new File(inExpected)), encoding), false);
 Lemmatizer dummyLemmatizer = new DummyLemmatizer(predictedSample);
 OutputStream stream = new ByteArrayOutputStream();
 LemmatizerEvaluationMonitor listener = new LemmaEvaluationErrorListener(stream);
 LemmatizerEvaluator evaluator = new LemmatizerEvaluator(dummyLemmatizer, listener);
 evaluator.evaluate(expectedSample);
 Assert.assertEquals(0.9877049180327869, evaluator.getWordAccuracy(), DELTA);
 Assert.assertNotSame(stream.toString().length(), 0);
}

private double trainAndEval(String lang, File trainFile, TrainingParameters params,
                  File evalFile) throws IOException {
 ConlluTagset tagset = ConlluTagset.X;
 ObjectStream<LemmaSample> trainSamples = new ConlluLemmaSampleStream(new ConlluStream(
   new MarkableFileInputStreamFactory(trainFile)), tagset);
 LemmatizerModel model = LemmatizerME.train(lang, trainSamples, params, new LemmatizerFactory());
 LemmatizerEvaluator evaluator = new LemmatizerEvaluator(new LemmatizerME(model));
 evaluator.evaluate(new ConlluLemmaSampleStream(new ConlluStream(
   new MarkableFileInputStreamFactory(evalFile)), tagset));
 return evaluator.getWordAccuracy();
}

LemmatizerEvaluator evaluator = new LemmatizerEvaluator(
  new opennlp.tools.lemmatizer.LemmatizerME(model),
  missclassifiedListener, reportListener);

LemmatizerEvaluator evaluator = new LemmatizerEvaluator(
  new opennlp.tools.lemmatizer.LemmatizerME(model),
  missclassifiedListener, reportListener);