Examples with CategoricalLabel org.jpmml.converter.CategoricalLabel used on opensource projects

Example 21 with CategoricalLabel

use of org.jpmml.converter.CategoricalLabel in project jpmml-r by jpmml.

the class RPartConverter method encodeClassification.

private TreeModel encodeClassification(RGenericVector frame, RIntegerVector rowNames, RVector<?> var, RIntegerVector n, int[][] splitInfo, RNumberVector<?> splits, RIntegerVector csplit, Schema schema) {
    RDoubleVector yval2 = frame.getDoubleElement("yval2");
    CategoricalLabel categoricalLabel = (CategoricalLabel) schema.getLabel();
    List<?> categories = categoricalLabel.getValues();
    boolean hasScoreDistribution = hasScoreDistribution();
    ScoreEncoder scoreEncoder = new ScoreEncoder() {

        private List<Integer> classes = null;

        private List<List<? extends Number>> recordCounts = null;

        {
            int rows = rowNames.size();
            int columns = 1 + (2 * categories.size()) + 1;
            List<Integer> classes = ValueUtil.asIntegers(FortranMatrixUtil.getColumn(yval2.getValues(), rows, columns, 0));
            this.classes = new ArrayList<>(classes);
            if (hasScoreDistribution) {
                this.recordCounts = new ArrayList<>();
                for (int i = 0; i < categories.size(); i++) {
                    List<? extends Number> recordCounts = FortranMatrixUtil.getColumn(yval2.getValues(), rows, columns, 1 + i);
                    this.recordCounts.add(new ArrayList<>(recordCounts));
                }
            }
        }

        @Override
        public Node encode(Node node, int offset) {
            Object score = categories.get(this.classes.get(offset) - 1);
            Integer recordCount = n.getValue(offset);
            node.setScore(score).setRecordCount(recordCount);
            if (hasScoreDistribution) {
                node = new ClassifierNode(node);
                List<ScoreDistribution> scoreDistributions = node.getScoreDistributions();
                for (int i = 0; i < categories.size(); i++) {
                    List<? extends Number> recordCounts = this.recordCounts.get(i);
                    ScoreDistribution scoreDistribution = new ScoreDistribution().setValue(categories.get(i)).setRecordCount(recordCounts.get(offset));
                    scoreDistributions.add(scoreDistribution);
                }
            }
            return node;
        }
    };
    Node root = encodeNode(True.INSTANCE, 1, rowNames, var, n, splitInfo, splits, csplit, scoreEncoder, schema);
    TreeModel treeModel = new TreeModel(MiningFunction.CLASSIFICATION, ModelUtil.createMiningSchema(schema.getLabel()), root);
    if (hasScoreDistribution) {
        treeModel.setOutput(ModelUtil.createProbabilityOutput(DataType.DOUBLE, categoricalLabel));
    }
    return configureTreeModel(treeModel);
}

Example 22 with CategoricalLabel

use of org.jpmml.converter.CategoricalLabel in project jpmml-r by jpmml.

the class RandomForestConverter method encodeClassification.

private MiningModel encodeClassification(RGenericVector forest, Schema schema) {
    RNumberVector<?> bestvar = forest.getNumericElement("bestvar");
    RNumberVector<?> treemap = forest.getNumericElement("treemap");
    RIntegerVector nodepred = forest.getIntegerElement("nodepred");
    RDoubleVector xbestsplit = forest.getDoubleElement("xbestsplit");
    RIntegerVector nrnodes = forest.getIntegerElement("nrnodes");
    RDoubleVector ntree = forest.getDoubleElement("ntree");
    int rows = nrnodes.asScalar();
    int columns = ValueUtil.asInt(ntree.asScalar());
    CategoricalLabel categoricalLabel = (CategoricalLabel) schema.getLabel();
    ScoreEncoder<Integer> scoreEncoder = new ScoreEncoder<Integer>() {

        @Override
        public Object encode(Integer value) {
            return categoricalLabel.getValue(value - 1);
        }
    };
    Schema segmentSchema = schema.toAnonymousSchema();
    List<TreeModel> treeModels = new ArrayList<>();
    for (int i = 0; i < columns; i++) {
        List<? extends Number> daughters = FortranMatrixUtil.getColumn(treemap.getValues(), 2 * rows, columns, i);
        TreeModel treeModel = encodeTreeModel(MiningFunction.CLASSIFICATION, scoreEncoder, FortranMatrixUtil.getColumn(daughters, rows, 2, 0), FortranMatrixUtil.getColumn(daughters, rows, 2, 1), FortranMatrixUtil.getColumn(nodepred.getValues(), rows, columns, i), FortranMatrixUtil.getColumn(bestvar.getValues(), rows, columns, i), FortranMatrixUtil.getColumn(xbestsplit.getValues(), rows, columns, i), segmentSchema);
        treeModels.add(treeModel);
    }
    MiningModel miningModel = new MiningModel(MiningFunction.CLASSIFICATION, ModelUtil.createMiningSchema(categoricalLabel)).setSegmentation(MiningModelUtil.createSegmentation(Segmentation.MultipleModelMethod.MAJORITY_VOTE, treeModels)).setOutput(ModelUtil.createProbabilityOutput(DataType.DOUBLE, categoricalLabel));
    return miningModel;
}

Example 23 with CategoricalLabel

use of org.jpmml.converter.CategoricalLabel in project jpmml-r by jpmml.

the class NNetConverter method encodeModel.

@Override
public Model encodeModel(Schema schema) {
    RGenericVector nnet = getObject();
    RDoubleVector n = nnet.getDoubleElement("n");
    RBooleanVector linout = nnet.getBooleanElement("linout", false);
    RBooleanVector softmax = nnet.getBooleanElement("softmax", false);
    RBooleanVector censored = nnet.getBooleanElement("censored", false);
    RDoubleVector wts = nnet.getDoubleElement("wts");
    RStringVector lev = nnet.getStringElement("lev", false);
    if (n.size() != 3) {
        throw new IllegalArgumentException();
    }
    Label label = schema.getLabel();
    List<? extends Feature> features = schema.getFeatures();
    MiningFunction miningFunction;
    if (lev == null) {
        if (linout != null && !linout.asScalar()) {
            throw new IllegalArgumentException();
        }
        miningFunction = MiningFunction.REGRESSION;
    } else {
        miningFunction = MiningFunction.CLASSIFICATION;
    }
    int nInput = ValueUtil.asInt(n.getValue(0));
    SchemaUtil.checkSize(nInput, features);
    NeuralInputs neuralInputs = NeuralNetworkUtil.createNeuralInputs(features, DataType.DOUBLE);
    int offset = 0;
    List<NeuralLayer> neuralLayers = new ArrayList<>();
    List<? extends NeuralEntity> entities = neuralInputs.getNeuralInputs();
    int nHidden = ValueUtil.asInt(n.getValue(1));
    if (nHidden > 0) {
        NeuralLayer neuralLayer = encodeNeuralLayer("hidden", nHidden, entities, wts, offset).setActivationFunction(NeuralNetwork.ActivationFunction.LOGISTIC);
        offset += (nHidden * (entities.size() + 1));
        neuralLayers.add(neuralLayer);
        entities = neuralLayer.getNeurons();
    }
    int nOutput = ValueUtil.asInt(n.getValue(2));
    if (nOutput == 1) {
        NeuralLayer neuralLayer = encodeNeuralLayer("output", nOutput, entities, wts, offset);
        offset += (nOutput * (entities.size() + 1));
        neuralLayers.add(neuralLayer);
        entities = neuralLayer.getNeurons();
        switch(miningFunction) {
            case REGRESSION:
                break;
            case CLASSIFICATION:
                {
                    List<NeuralLayer> transformationNeuralLayers = NeuralNetworkUtil.createBinaryLogisticTransformation(Iterables.getOnlyElement(entities));
                    neuralLayers.addAll(transformationNeuralLayers);
                    neuralLayer = Iterables.getLast(transformationNeuralLayers);
                    entities = neuralLayer.getNeurons();
                }
                break;
        }
    } else if (nOutput > 1) {
        NeuralLayer neuralLayer = encodeNeuralLayer("output", nOutput, entities, wts, offset);
        if (softmax != null && softmax.asScalar()) {
            if (censored != null && censored.asScalar()) {
                throw new IllegalArgumentException();
            }
            neuralLayer.setNormalizationMethod(NeuralNetwork.NormalizationMethod.SOFTMAX);
        }
        offset += (nOutput * (entities.size() + 1));
        neuralLayers.add(neuralLayer);
        entities = neuralLayer.getNeurons();
    } else {
        throw new IllegalArgumentException();
    }
    NeuralNetwork neuralNetwork = new NeuralNetwork(miningFunction, NeuralNetwork.ActivationFunction.IDENTITY, ModelUtil.createMiningSchema(label), neuralInputs, neuralLayers);
    switch(miningFunction) {
        case REGRESSION:
            neuralNetwork.setNeuralOutputs(NeuralNetworkUtil.createRegressionNeuralOutputs(entities, (ContinuousLabel) label));
            break;
        case CLASSIFICATION:
            neuralNetwork.setNeuralOutputs(NeuralNetworkUtil.createClassificationNeuralOutputs(entities, (CategoricalLabel) label)).setOutput(ModelUtil.createProbabilityOutput(DataType.DOUBLE, (CategoricalLabel) label));
            break;
    }
    return neuralNetwork;
}

Example 24 with CategoricalLabel

use of org.jpmml.converter.CategoricalLabel in project jpmml-r by jpmml.

the class RExpEncoder method setLabel.

public void setLabel(DataField dataField) {
    Label label;
    OpType opType = dataField.requireOpType();
    switch(opType) {
        case CATEGORICAL:
            label = new CategoricalLabel(dataField);
            break;
        case CONTINUOUS:
            label = new ContinuousLabel(dataField);
            break;
        default:
            throw new IllegalArgumentException();
    }
    setLabel(label);
}

Example 25 with CategoricalLabel

use of org.jpmml.converter.CategoricalLabel in project jpmml-r by jpmml.

the class RangerConverter method encodeProbabilityForest.

private MiningModel encodeProbabilityForest(RGenericVector forest, Schema schema) {
    RStringVector levels = forest.getStringElement("levels");
    CategoricalLabel categoricalLabel = (CategoricalLabel) schema.getLabel();
    ScoreEncoder scoreEncoder = new ScoreEncoder() {

        @Override
        public Node encode(Node node, Number splitValue, RNumberVector<?> terminalClassCount) {
            if (splitValue.doubleValue() != 0d || (terminalClassCount == null || terminalClassCount.size() != levels.size())) {
                throw new IllegalArgumentException();
            }
            node = new ClassifierNode(node);
            List<ScoreDistribution> scoreDistributions = node.getScoreDistributions();
            Number maxProbability = null;
            for (int i = 0; i < terminalClassCount.size(); i++) {
                String value = levels.getValue(i);
                Number probability = terminalClassCount.getValue(i);
                if (maxProbability == null || ((Comparable) maxProbability).compareTo(probability) < 0) {
                    node.setScore(value);
                    maxProbability = probability;
                }
                ScoreDistribution scoreDistribution = new ScoreDistribution(value, probability);
                scoreDistributions.add(scoreDistribution);
            }
            return node;
        }
    };
    List<TreeModel> treeModels = encodeForest(forest, MiningFunction.CLASSIFICATION, scoreEncoder, schema);
    MiningModel miningModel = new MiningModel(MiningFunction.CLASSIFICATION, ModelUtil.createMiningSchema(categoricalLabel)).setSegmentation(MiningModelUtil.createSegmentation(Segmentation.MultipleModelMethod.AVERAGE, treeModels)).setOutput(ModelUtil.createProbabilityOutput(DataType.DOUBLE, categoricalLabel));
    return miningModel;
}