Example usage for org.apache.commons.math.linear RealMatrix getColumnVector

Introduction

In this page you can find the example usage for org.apache.commons.math.linear RealMatrix getColumnVector.

Prototype

RealVector getColumnVector(int column) throws MatrixIndexException;

Source Link

Document

Returns the entries in column number column as a vector.

Usage

From source file:de.mpicbg.knime.hcs.base.nodes.preproc.OutlierRemoval.java

@Override
protected BufferedDataTable[] execute(BufferedDataTable[] inData, ExecutionContext exec) throws Exception {

    BufferedDataTable input = inData[0];
    DataTableSpec inputSpec = input.getDataTableSpec();

    // Get the parameter and make sure there all double value columns
    List<Attribute> parameter = new ArrayList<Attribute>();
    for (String item : parameterNames.getIncludeList()) {
        Attribute attribute = new InputTableAttribute(item, input);
        if (attribute.getType().isCompatible(DoubleValue.class)) {
            parameter.add(attribute);//  w w  w.  ja  va 2  s  . co m
        } else {
            logger.warn("The parameter '" + attribute.getName()
                    + "' will not be considered for outlier removal, since it is not compatible to double.");
        }
    }

    // Get the groups defined by the nominal column.
    Attribute groupingAttribute = new InputTableAttribute(this.groupingColumn.getStringValue(), input);
    Map<Object, List<DataRow>> subsets = AttributeUtils.splitRowsGeneric(input, groupingAttribute);

    // Initialize
    BufferedDataContainer keepContainer = exec.createDataContainer(inputSpec);
    BufferedDataContainer discartContainer = exec.createDataContainer(inputSpec);
    int S = subsets.size();
    int s = 1;

    // Outlier analysis for each subset
    for (Object key : subsets.keySet()) {

        // Get the subset having all constraints in common
        List<DataRow> rowSubset = subsets.get(key);

        // Get the valid values
        RealMatrix data = extractMatrix(rowSubset, parameter);

        int N = data.getColumnDimension();
        int M = data.getRowDimension();
        if (M == 0) {
            logger.warn("The group '" + key + "' has no valid values and will be removed entirely'");
        } else {

            // Determine upper and lower outlier bounds
            double[] lowerBound = new double[N];
            double[] upperBound = new double[N];
            if (method.getStringValue().equals("Boxplot")) {
                for (int c = 0; c < N; ++c) {
                    RealVector vect = data.getColumnVector(c);
                    DescriptiveStatistics stats = new DescriptiveStatistics();
                    for (double value : vect.getData()) {
                        stats.addValue(value);
                    }
                    double lowerQuantile = stats.getPercentile(25);
                    double upperQuantile = stats.getPercentile(85);
                    double whisker = factor.getDoubleValue() * Math.abs(lowerQuantile - upperQuantile);
                    lowerBound[c] = lowerQuantile - whisker;
                    upperBound[c] = upperQuantile + whisker;
                }
            } else {
                for (int c = 0; c < N; ++c) {
                    RealVector vect = data.getColumnVector(c);
                    double mean = StatUtils.mean(vect.getData());
                    double sd = Math.sqrt(StatUtils.variance(vect.getData()));
                    lowerBound[c] = mean - factor.getDoubleValue() * sd;
                    upperBound[c] = mean + factor.getDoubleValue() * sd;
                }
            }

            // Remove The outlier
            if (rule.getBooleanValue()) { // The row is only discarted if the row is an outlier in all parameter.
                for (DataRow row : rowSubset) {
                    int c = 0;
                    for (Attribute column : parameter) {

                        DataCell valueCell = row.getCell(((InputTableAttribute) column).getColumnIndex());

                        // a missing value will be treated as data point inside the bounds
                        if (valueCell.isMissing()) {
                            continue;
                        }

                        Double value = ((DoubleValue) valueCell).getDoubleValue();
                        if ((value != null) && (lowerBound[c] <= value) && (value <= upperBound[c])) {
                            break;
                        } else {
                            c++;
                        }
                    }
                    if (c != N) {
                        keepContainer.addRowToTable(row);
                    } else {
                        discartContainer.addRowToTable(row);
                    }
                }
            } else { // The row is discarted if it has a outlier for at least one parameter.
                for (DataRow row : rowSubset) {
                    int c = 0;
                    for (Attribute column : parameter) {

                        DataCell valueCell = row.getCell(((InputTableAttribute) column).getColumnIndex());

                        // a missing value will be treated as data point inside the bounds
                        if (valueCell.isMissing()) {
                            c++;
                            continue;
                        }

                        Double value = ((DoubleValue) valueCell).getDoubleValue();
                        if ((value != null) && (lowerBound[c] <= value) && (value <= upperBound[c])) {
                            c++;
                        } else {
                            break;
                        }
                    }
                    if (c == N) {
                        keepContainer.addRowToTable(row);
                    } else {
                        discartContainer.addRowToTable(row);
                    }
                }
            }
        }

        BufTableUtils.updateProgress(exec, s++, S);

    }

    keepContainer.close();
    discartContainer.close();
    return new BufferedDataTable[] { keepContainer.getTable(), discartContainer.getTable() };
}

---