[IGNITE-12269] [ML] The method printTree is corrupted on the example

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 2.8
• Fix Version/s: 2.8
• Component/s: ml
• Release Note: The method printTree is corrupted on the example
• Ignite Flags: Release Notes Required
• Assignee: Alexey Zinoviev
• Reporter: Alexey Zinoviev
• Created: 08/Oct/19 12:55
• Updated: 14/Jan/20 09:07
• Resolved: 09/Oct/19 11:53

Description

This example

package org.apache.ignite.examples.ml.tutorial;

import java.io.FileNotFoundException;
import org.apache.ignite.Ignite;
import org.apache.ignite.IgniteCache;
import org.apache.ignite.Ignition;
import org.apache.ignite.ml.composition.ModelsComposition;
import org.apache.ignite.ml.composition.boosting.convergence.mean.MeanAbsValueConvergenceCheckerFactory;
import org.apache.ignite.ml.composition.boosting.convergence.median.MedianOfMedianConvergenceCheckerFactory;
import org.apache.ignite.ml.dataset.feature.extractor.Vectorizer;
import org.apache.ignite.ml.dataset.feature.extractor.impl.DummyVectorizer;
import org.apache.ignite.ml.math.primitives.vector.Vector;
import org.apache.ignite.ml.preprocessing.Preprocessor;
import org.apache.ignite.ml.preprocessing.encoding.EncoderTrainer;
import org.apache.ignite.ml.preprocessing.encoding.EncoderType;
import org.apache.ignite.ml.preprocessing.imputing.ImputerTrainer;
import org.apache.ignite.ml.preprocessing.minmaxscaling.MinMaxScalerTrainer;
import org.apache.ignite.ml.preprocessing.normalization.NormalizationTrainer;
import org.apache.ignite.ml.selection.scoring.evaluator.Evaluator;
import org.apache.ignite.ml.selection.scoring.metric.MetricName;
import org.apache.ignite.ml.selection.split.TrainTestDatasetSplitter;
import org.apache.ignite.ml.selection.split.TrainTestSplit;
import org.apache.ignite.ml.trainers.DatasetTrainer;
import org.apache.ignite.ml.tree.boosting.GDBBinaryClassifierOnTreesTrainer;

/**
 * {@link MinMaxScalerTrainer} and {@link NormalizationTrainer} are used in this example due to different values
 * distribution in columns and rows.
 * <p>
 * Code in this example launches Ignite grid and fills the cache with test data (based on Titanic passengers data).</p>
 * <p>
 * After that it defines preprocessors that extract features from an upstream data and perform other desired changes
 * over the extracted data, including the scaling.</p>
 * <p>
 * Then, it trains the model based on the processed data using decision tree classification.</p>
 * <p>
 * Finally, this example uses {@link Evaluator} functionality to compute metrics from predictions.</p>
 */
public class Step_11_Boosting {
    /**
     * Run example.
     */
    public static void main(String[] args) {
        System.out.println();
        System.out.println(">>> Tutorial step 11 (Boosting) example started.");

        try (Ignite ignite = Ignition.start("examples/config/example-ignite.xml")) {
            try {
                IgniteCache<Integer, Vector> dataCache = TitanicUtils.readPassengers(ignite);

                // Extracts "pclass", "sibsp", "parch", "sex", "embarked", "age", "fare".
                final Vectorizer<Integer, Vector, Integer, Double> vectorizer
                    = new DummyVectorizer<Integer>(0, 3, 4, 5, 6, 8, 10).labeled(1);

                TrainTestSplit<Integer, Vector> split = new TrainTestDatasetSplitter<Integer, Vector>()
                    .split(0.75);

                Preprocessor<Integer, Vector> strEncoderPreprocessor = new EncoderTrainer<Integer, Vector>()
                    .withEncoderType(EncoderType.STRING_ENCODER)
                    .withEncodedFeature(1)
                    .withEncodedFeature(6) // <--- Changed index here.
                    .fit(ignite,
                        dataCache,
                        vectorizer
                    );

                Preprocessor<Integer, Vector> imputingPreprocessor = new ImputerTrainer<Integer, Vector>()
                    .fit(ignite,
                        dataCache,
                        strEncoderPreprocessor
                    );

                Preprocessor<Integer, Vector> minMaxScalerPreprocessor = new MinMaxScalerTrainer<Integer, Vector>()
                    .fit(
                        ignite,
                        dataCache,
                        imputingPreprocessor
                    );

                Preprocessor<Integer, Vector> normalizationPreprocessor = new NormalizationTrainer<Integer, Vector>()
                    .withP(1)
                    .fit(
                        ignite,
                        dataCache,
                        minMaxScalerPreprocessor
                    );

                // Create classification trainer.
                DatasetTrainer<ModelsComposition, Double> trainer = new GDBBinaryClassifierOnTreesTrainer(0.01, 1000, 10, 0.)
                    .withCheckConvergenceStgyFactory(new MedianOfMedianConvergenceCheckerFactory(0.01));

                // Train decision tree model.
                ModelsComposition mdl = trainer.fit(
                    ignite,
                    dataCache,
                    split.getTrainFilter(),
                    normalizationPreprocessor
                );

                System.out.println("\n>>> Trained model: " + mdl);

                double accuracy = Evaluator.evaluate(
                    dataCache,
                    split.getTestFilter(),
                    mdl,
                    normalizationPreprocessor,
                    MetricName.ACCURACY
                );

                System.out.println("\n>>> Accuracy " + accuracy);
                System.out.println("\n>>> Test Error " + (1 - accuracy));

                System.out.println(">>> Tutorial step 11 (Boosting) example completed.");
            }
            catch (FileNotFoundException e) {
                e.printStackTrace();
            }
        }
        finally {
            System.out.flush();
        }
    }
}

is ruined with

Exception in thread "main" java.lang.IllegalArgumentExceptionException in thread "main" java.lang.IllegalArgumentException at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:105) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:100) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:102) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:100) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:100) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:100) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:102) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:102) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:100) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:100) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:102) at org.apache.ignite.ml.tree.DecisionTree.printTree(DecisionTree.java:328) at org.apache.ignite.ml.tree.DecisionTreeConditionalNode.toString(DecisionTreeConditionalNode.java:123) at org.apache.ignite.ml.tree.DecisionTreeConditionalNode.toString(DecisionTreeConditionalNode.java:118) at java.lang.String.valueOf(String.java:2994) at java.lang.StringBuilder.append(StringBuilder.java:131) at org.apache.ignite.ml.util.ModelTrace.lambda$fieldToString$1(ModelTrace.java:122) at java.util.stream.ReferencePipeline$3$1.accept(ReferencePipeline.java:193) at java.util.ArrayList$ArrayListSpliterator.forEachRemaining(ArrayList.java:1374) at java.util.stream.AbstractPipeline.copyInto(AbstractPipeline.java:481) at java.util.stream.AbstractPipeline.wrapAndCopyInto(AbstractPipeline.java:471) at java.util.stream.ReduceOps$ReduceOp.evaluateSequential(ReduceOps.java:708) at java.util.stream.AbstractPipeline.evaluate(AbstractPipeline.java:234) at java.util.stream.ReferencePipeline.collect(ReferencePipeline.java:499) at org.apache.ignite.ml.util.ModelTrace.fieldToString(ModelTrace.java:123) at org.apache.ignite.ml.util.ModelTrace.lambda$toString$0(ModelTrace.java:97) at java.util.stream.ReferencePipeline$3$1.accept(ReferencePipeline.java:193) at java.util.ArrayList$ArrayListSpliterator.forEachRemaining(ArrayList.java:1374) at java.util.stream.AbstractPipeline.copyInto(AbstractPipeline.java:481) at java.util.stream.AbstractPipeline.wrapAndCopyInto(AbstractPipeline.java:471) at java.util.stream.ReduceOps$ReduceOp.evaluateSequential(ReduceOps.java:708) at java.util.stream.AbstractPipeline.evaluate(AbstractPipeline.java:234) at java.util.stream.ReferencePipeline.collect(ReferencePipeline.java:499) at org.apache.ignite.ml.util.ModelTrace.toString(ModelTrace.java:98) at org.apache.ignite.ml.composition.ModelsComposition.toString(ModelsComposition.java:101) at org.apache.ignite.ml.composition.ModelsComposition.toString(ModelsComposition.java:93) at java.lang.String.valueOf(String.java:2994) at java.lang.StringBuilder.append(StringBuilder.java:131) at org.apache.ignite.examples.ml.tutorial.Step_11_Boosting.main(Step_11_Boosting.java:117)

 

 

Issue Links

• relates to: IGNITE-12274 [ML] DecisionTree works incorrectly if maxDeep > amount of features
• links to: GitHub Pull Request #6958