Examples with Regressor org.tribuo.regression.Regressor used on opensource projects

Example 1 with Regressor

use of org.tribuo.regression.Regressor in project ml-commons by opensearch-project.

the class TribuoUtil method generateDatasetWithTarget.

/**
 * Generate tribuo dataset from data frame with target.
 * @param dataFrame features data
 * @param outputFactory the tribuo output factory
 * @param desc description for tribuo provenance
 * @param outputType the tribuo output type
 * @param target target name
 * @return tribuo dataset
 */
public static <T extends Output<T>> MutableDataset<T> generateDatasetWithTarget(DataFrame dataFrame, OutputFactory<T> outputFactory, String desc, TribuoOutputType outputType, String target) {
    if (StringUtils.isEmpty(target)) {
        throw new IllegalArgumentException("Empty target when generating dataset from data frame.");
    }
    List<Example<T>> dataset = new ArrayList<>();
    Tuple<String[], double[][]> featureNamesValues = transformDataFrame(dataFrame);
    int targetIndex = -1;
    for (int i = 0; i < featureNamesValues.v1().length; ++i) {
        if (featureNamesValues.v1()[i].equals(target)) {
            targetIndex = i;
            break;
        }
    }
    if (targetIndex == -1) {
        throw new IllegalArgumentException("No matched target when generating dataset from data frame.");
    }
    ArrayExample<T> example;
    final int finalTargetIndex = targetIndex;
    String[] featureNames = IntStream.range(0, featureNamesValues.v1().length).filter(e -> e != finalTargetIndex).mapToObj(e -> featureNamesValues.v1()[e]).toArray(String[]::new);
    for (int i = 0; i < dataFrame.size(); ++i) {
        switch(outputType) {
            case REGRESSOR:
                final int finalI = i;
                double targetValue = featureNamesValues.v2()[finalI][finalTargetIndex];
                double[] featureValues = IntStream.range(0, featureNamesValues.v2()[i].length).filter(e -> e != finalTargetIndex).mapToDouble(e -> featureNamesValues.v2()[finalI][e]).toArray();
                example = new ArrayExample<>((T) new Regressor(target, targetValue), featureNames, featureValues);
                break;
            default:
                throw new IllegalArgumentException("unknown type:" + outputType);
        }
        dataset.add(example);
    }
    SimpleDataSourceProvenance provenance = new SimpleDataSourceProvenance(desc, outputFactory);
    return new MutableDataset<>(new ListDataSource<>(dataset, outputFactory, provenance));
}

Example 2 with Regressor

use of org.tribuo.regression.Regressor in project ml-commons by opensearch-project.

the class TribuoUtil method generateDataset.

/**
 * Generate tribuo dataset from data frame.
 * @param dataFrame features data
 * @param outputFactory the tribuo output factory
 * @param desc description for tribuo provenance
 * @param outputType the tribuo output type
 * @return tribuo dataset
 */
public static <T extends Output<T>> MutableDataset<T> generateDataset(DataFrame dataFrame, OutputFactory<T> outputFactory, String desc, TribuoOutputType outputType) {
    List<Example<T>> dataset = new ArrayList<>();
    Tuple<String[], double[][]> featureNamesValues = transformDataFrame(dataFrame);
    ArrayExample<T> example;
    for (int i = 0; i < dataFrame.size(); ++i) {
        switch(outputType) {
            case CLUSTERID:
                example = new ArrayExample<>((T) new ClusterID(ClusterID.UNASSIGNED), featureNamesValues.v1(), featureNamesValues.v2()[i]);
                break;
            case REGRESSOR:
                // Create single dimension tribuo regressor with name DIM-0 and value double NaN.
                example = new ArrayExample<>((T) new Regressor("DIM-0", Double.NaN), featureNamesValues.v1(), featureNamesValues.v2()[i]);
                break;
            case ANOMALY_DETECTION_LIBSVM:
                // Why we set default event type as EXPECTED(non-anomalous)
                // 1. For training data, Tribuo LibSVMAnomalyTrainer only supports EXPECTED events at training time.
                // 2. For prediction data, we treat the data as non-anomalous by default as Tribuo lib don't accept UNKNOWN type.
                Event.EventType defaultEventType = Event.EventType.EXPECTED;
                // TODO: support anomaly labels to evaluate prediction result
                example = new ArrayExample<>((T) new Event(defaultEventType), featureNamesValues.v1(), featureNamesValues.v2()[i]);
                break;
            default:
                throw new IllegalArgumentException("unknown type:" + outputType);
        }
        dataset.add(example);
    }
    SimpleDataSourceProvenance provenance = new SimpleDataSourceProvenance(desc, outputFactory);
    return new MutableDataset<>(new ListDataSource<>(dataset, outputFactory, provenance));
}

Example 3 with Regressor

use of org.tribuo.regression.Regressor in project ml-commons by opensearch-project.

the class LinearRegression method predict.

@Override
public MLOutput predict(DataFrame dataFrame, Model model) {
    if (model == null) {
        throw new IllegalArgumentException("No model found for linear regression prediction.");
    }
    org.tribuo.Model<Regressor> regressionModel = (org.tribuo.Model<Regressor>) ModelSerDeSer.deserialize(model.getContent());
    MutableDataset<Regressor> predictionDataset = TribuoUtil.generateDataset(dataFrame, new RegressionFactory(), "Linear regression prediction data from opensearch", TribuoOutputType.REGRESSOR);
    List<Prediction<Regressor>> predictions = regressionModel.predict(predictionDataset);
    List<Map<String, Object>> listPrediction = new ArrayList<>();
    predictions.forEach(e -> listPrediction.add(Collections.singletonMap(e.getOutput().getNames()[0], e.getOutput().getValues()[0])));
    return MLPredictionOutput.builder().predictionResult(DataFrameBuilder.load(listPrediction)).build();
}

Example 4 with Regressor

use of org.tribuo.regression.Regressor in project ml-commons by opensearch-project.

the class TribuoUtilTest method generateDatasetWithTarget.

@SuppressWarnings("unchecked")
@Test
public void generateDatasetWithTarget() {
    MutableDataset<Regressor> dataset = TribuoUtil.generateDatasetWithTarget(dataFrame, new RegressionFactory(), "test", TribuoOutputType.REGRESSOR, "f2");
    List<Example<Regressor>> examples = dataset.getData();
    Assert.assertEquals(rawData.length, examples.size());
    for (int i = 0; i < rawData.length; ++i) {
        ArrayExample arrayExample = (ArrayExample) examples.get(i);
        Iterator<Feature> iterator = arrayExample.iterator();
        int idx = 1;
        while (iterator.hasNext()) {
            Feature feature = iterator.next();
            Assert.assertEquals("f" + idx, feature.getName());
            Assert.assertEquals(i + idx / 10.0, feature.getValue(), 0.01);
            ++idx;
        }
    }
}

Example 5 with Regressor

use of org.tribuo.regression.Regressor in project ml-commons by opensearch-project.

the class LinearRegression method train.

@Override
public Model train(DataFrame dataFrame) {
    MutableDataset<Regressor> trainDataset = TribuoUtil.generateDatasetWithTarget(dataFrame, new RegressionFactory(), "Linear regression training data from opensearch", TribuoOutputType.REGRESSOR, parameters.getTarget());
    Integer epochs = Optional.ofNullable(parameters.getEpochs()).orElse(DEFAULT_EPOCHS);
    LinearSGDTrainer linearSGDTrainer = new LinearSGDTrainer(objective, optimiser, epochs, DEFAULT_INTERVAL, DEFAULT_BATCH_SIZE, seed);
    org.tribuo.Model<Regressor> regressionModel = linearSGDTrainer.train(trainDataset);
    Model model = new Model();
    model.setName(FunctionName.LINEAR_REGRESSION.name());
    model.setVersion(1);
    model.setContent(ModelSerDeSer.serialize(regressionModel));
    return model;
}