Examples with LogisticRegressionSGDTrainer org.apache.ignite.ml.regressions.logistic.LogisticRegressionSGDTrainer used on opensource projects

Example 11 with LogisticRegressionSGDTrainer

use of org.apache.ignite.ml.regressions.logistic.LogisticRegressionSGDTrainer in project ignite by apache.

the class CrossValidationTest method testRandomSearch.

/**
 */
@Test
public void testRandomSearch() {
    Map<Integer, double[]> data = new HashMap<>();
    for (int i = 0; i < twoLinearlySeparableClasses.length; i++) data.put(i, twoLinearlySeparableClasses[i]);
    LogisticRegressionSGDTrainer trainer = new LogisticRegressionSGDTrainer().withUpdatesStgy(new UpdatesStrategy<>(new SimpleGDUpdateCalculator(0.2), SimpleGDParameterUpdate.SUM_LOCAL, SimpleGDParameterUpdate.AVG)).withMaxIterations(100000).withLocIterations(100).withBatchSize(14).withSeed(123L);
    Vectorizer<Integer, double[], Integer, Double> vectorizer = new DoubleArrayVectorizer<Integer>().labeled(Vectorizer.LabelCoordinate.FIRST);
    ParamGrid paramGrid = new ParamGrid().withParameterSearchStrategy(new RandomStrategy().withMaxTries(10).withSeed(1234L).withSatisfactoryFitness(0.9)).addHyperParam("maxIterations", trainer::withMaxIterations, new Double[] { 10.0, 100.0, 1000.0, 10000.0 }).addHyperParam("locIterations", trainer::withLocIterations, new Double[] { 10.0, 100.0, 1000.0, 10000.0 }).addHyperParam("batchSize", trainer::withBatchSize, new Double[] { 1.0, 2.0, 4.0, 8.0, 16.0 });
    DebugCrossValidation<LogisticRegressionModel, Integer, double[]> scoreCalculator = (DebugCrossValidation<LogisticRegressionModel, Integer, double[]>) new DebugCrossValidation<LogisticRegressionModel, Integer, double[]>().withUpstreamMap(data).withAmountOfParts(1).withTrainer(trainer).withMetric(MetricName.ACCURACY).withPreprocessor(vectorizer).withAmountOfFolds(4).isRunningOnPipeline(false).withParamGrid(paramGrid);
    CrossValidationResult crossValidationRes = scoreCalculator.tuneHyperParameters();
    assertEquals(crossValidationRes.getBestAvgScore(), 0.9343858500738256, 1e-6);
    assertEquals(crossValidationRes.getScoringBoard().size(), 10);
}

Example 12 with LogisticRegressionSGDTrainer

use of org.apache.ignite.ml.regressions.logistic.LogisticRegressionSGDTrainer in project ignite by apache.

the class OneVsRestTrainerTest method testTrainWithTheLinearlySeparableCase.

/**
 * Test trainer on 2 linearly separable sets.
 */
@Test
public void testTrainWithTheLinearlySeparableCase() {
    Map<Integer, double[]> cacheMock = new HashMap<>();
    for (int i = 0; i < twoLinearlySeparableClasses.length; i++) cacheMock.put(i, twoLinearlySeparableClasses[i]);
    LogisticRegressionSGDTrainer binaryTrainer = new LogisticRegressionSGDTrainer().withUpdatesStgy(new UpdatesStrategy<>(new SimpleGDUpdateCalculator(0.2), SimpleGDParameterUpdate.SUM_LOCAL, SimpleGDParameterUpdate.AVG)).withMaxIterations(1000).withLocIterations(10).withBatchSize(100).withSeed(123L);
    OneVsRestTrainer<LogisticRegressionModel> trainer = new OneVsRestTrainer<>(binaryTrainer);
    MultiClassModel mdl = trainer.fit(cacheMock, parts, new DoubleArrayVectorizer<Integer>().labeled(Vectorizer.LabelCoordinate.FIRST));
    Assert.assertTrue(!mdl.toString().isEmpty());
    Assert.assertTrue(!mdl.toString(true).isEmpty());
    Assert.assertTrue(!mdl.toString(false).isEmpty());
    TestUtils.assertEquals(1, mdl.predict(VectorUtils.of(-100, 0)), PRECISION);
    TestUtils.assertEquals(0, mdl.predict(VectorUtils.of(100, 0)), PRECISION);
}

Example 13 with LogisticRegressionSGDTrainer

use of org.apache.ignite.ml.regressions.logistic.LogisticRegressionSGDTrainer in project ignite by apache.

the class OneVsRestTrainerTest method testUpdate.

/**
 */
@Test
public void testUpdate() {
    Map<Integer, double[]> cacheMock = new HashMap<>();
    for (int i = 0; i < twoLinearlySeparableClasses.length; i++) cacheMock.put(i, twoLinearlySeparableClasses[i]);
    LogisticRegressionSGDTrainer binaryTrainer = new LogisticRegressionSGDTrainer().withUpdatesStgy(new UpdatesStrategy<>(new SimpleGDUpdateCalculator(0.2), SimpleGDParameterUpdate.SUM_LOCAL, SimpleGDParameterUpdate.AVG)).withMaxIterations(1000).withLocIterations(10).withBatchSize(100).withSeed(123L);
    OneVsRestTrainer<LogisticRegressionModel> trainer = new OneVsRestTrainer<>(binaryTrainer);
    Vectorizer<Integer, double[], Integer, Double> vectorizer = new DoubleArrayVectorizer<Integer>().labeled(Vectorizer.LabelCoordinate.FIRST);
    MultiClassModel originalMdl = trainer.fit(cacheMock, parts, vectorizer);
    MultiClassModel updatedOnSameDS = trainer.update(originalMdl, cacheMock, parts, vectorizer);
    MultiClassModel updatedOnEmptyDS = trainer.update(originalMdl, new HashMap<>(), parts, vectorizer);
    List<Vector> vectors = Arrays.asList(VectorUtils.of(-100, 0), VectorUtils.of(100, 0));
    for (Vector vec : vectors) {
        TestUtils.assertEquals(originalMdl.predict(vec), updatedOnSameDS.predict(vec), PRECISION);
        TestUtils.assertEquals(originalMdl.predict(vec), updatedOnEmptyDS.predict(vec), PRECISION);
    }
}

Example 14 with LogisticRegressionSGDTrainer

use of org.apache.ignite.ml.regressions.logistic.LogisticRegressionSGDTrainer in project ignite by apache.

the class MLDeployingTest method fitAndTestModel.

/**
 */
private void fitAndTestModel(CacheBasedDatasetBuilder<Integer, Vector> datasetBuilder, Preprocessor<Integer, Vector> preprocessor) {
    LogisticRegressionSGDTrainer trainer = new LogisticRegressionSGDTrainer();
    LogisticRegressionModel mdl = trainer.fit(datasetBuilder, preprocessor);
    // For this case any answer is valid.
    assertEquals(0., mdl.predict(VectorUtils.of(0., 0.)), 1.);
}