Examples with ElasticNetLinearRegOptimizer edu.neu.ccs.pyramid.regression.linear_regression.ElasticNetLinearRegOptimizer used on opensource projects

Example 1 with ElasticNetLinearRegOptimizer

use of edu.neu.ccs.pyramid.regression.linear_regression.ElasticNetLinearRegOptimizer in project pyramid by cheng-li.

the class LinearRegElasticNet method main.

public static void main(String[] args) throws Exception {
    if (args.length != 1) {
        throw new IllegalArgumentException("Please specify a properties file.");
    }
    Config config = new Config(args[0]);
    System.out.println(config);
    String output = config.getString("output.folder");
    new File(output).mkdirs();
    String sparsity = config.getString("featureMatrix.sparsity").toLowerCase();
    DataSetType dataSetType = null;
    switch(sparsity) {
        case "dense":
            dataSetType = DataSetType.REG_DENSE;
            break;
        case "sparse":
            dataSetType = DataSetType.REG_SPARSE;
            break;
        default:
            throw new IllegalArgumentException("featureMatrix.sparsity can be either dense or sparse");
    }
    RegDataSet trainSet = TRECFormat.loadRegDataSet(config.getString("input.trainSet"), dataSetType, true);
    RegDataSet testSet = TRECFormat.loadRegDataSet(config.getString("input.testSet"), dataSetType, true);
    LinearRegression linearRegression = new LinearRegression(trainSet.getNumFeatures());
    ElasticNetLinearRegOptimizer optimizer = new ElasticNetLinearRegOptimizer(linearRegression, trainSet);
    optimizer.setRegularization(config.getDouble("regularization"));
    optimizer.setL1Ratio(config.getDouble("l1Ratio"));
    System.out.println("before training");
    System.out.println("training set RMSE = " + RMSE.rmse(linearRegression, trainSet));
    System.out.println("test set RMSE = " + RMSE.rmse(linearRegression, testSet));
    System.out.println("start training");
    StopWatch stopWatch = new StopWatch();
    stopWatch.start();
    optimizer.optimize();
    System.out.println("training done");
    System.out.println("time spent on training = " + stopWatch);
    System.out.println("after training");
    System.out.println("training set RMSE = " + RMSE.rmse(linearRegression, trainSet));
    System.out.println("test set RMSE = " + RMSE.rmse(linearRegression, testSet));
    System.out.println("number of non-zeros weights in linear regression (not including bias) = " + linearRegression.getWeights().getWeightsWithoutBias().getNumNonZeroElements());
    List<Pair<Integer, Double>> sorted = new ArrayList<>();
    for (Vector.Element element : linearRegression.getWeights().getWeightsWithoutBias().nonZeroes()) {
        sorted.add(new Pair<>(element.index(), element.get()));
    }
    Comparator<Pair<Integer, Double>> comparatorByIndex = Comparator.comparing(pair -> pair.getFirst());
    sorted = sorted.stream().sorted(comparatorByIndex).collect(Collectors.toList());
    StringBuilder sb1 = new StringBuilder();
    for (Pair<Integer, Double> pair : sorted) {
        int index = pair.getFirst();
        sb1.append(index).append("(").append(trainSet.getFeatureList().get(index).getName()).append(")").append(":").append(pair.getSecond()).append("\n");
    }
    FileUtils.writeStringToFile(new File(output, "features_sorted_by_indices.txt"), sb1.toString());
    System.out.println("all selected features (sorted by indices) are saved to " + new File(output, "features_sorted_by_indices.txt").getAbsolutePath());
    Comparator<Pair<Integer, Double>> comparator = Comparator.comparing(pair -> Math.abs(pair.getSecond()));
    sorted = sorted.stream().sorted(comparator.reversed()).collect(Collectors.toList());
    StringBuilder sb = new StringBuilder();
    for (Pair<Integer, Double> pair : sorted) {
        int index = pair.getFirst();
        sb.append(index).append("(").append(trainSet.getFeatureList().get(index).getName()).append(")").append(":").append(pair.getSecond()).append("\n");
    }
    FileUtils.writeStringToFile(new File(output, "features_sorted_by_weights.txt"), sb.toString());
    System.out.println("all selected features (sorted by absolute weights) are saved to " + new File(output, "features_sorted_by_weights.txt").getAbsolutePath());
    File reportFile = new File(output, "test_predictions.txt");
    report(linearRegression, testSet, reportFile);
    System.out.println("predictions on the test set are written to " + reportFile.getAbsolutePath());
}