use of org.knime.core.data.DataColumnSpecCreator in project knime-core by knime.
the class PolyRegLearnerNodeModel method getCellFactory.
private CellFactory getCellFactory(final int dependentIndex) {
final int degree = m_settings.getDegree();
return new CellFactory() {
@Override
public DataCell[] getCells(final DataRow row) {
double sum = m_betas[0];
int betaCount = 1;
double y = 0;
for (int col = 0; col < row.getNumCells(); col++) {
if ((col != dependentIndex) && m_colSelected[col]) {
final double value = ((DoubleValue) row.getCell(col)).getDoubleValue();
double poly = 1;
for (int d = 1; d <= degree; d++) {
poly *= value;
sum += m_betas[betaCount++] * poly;
}
} else if (col == dependentIndex) {
y = ((DoubleValue) row.getCell(col)).getDoubleValue();
}
}
double err = Math.abs(sum - y);
m_squaredError += err * err;
return new DataCell[] { new DoubleCell(sum), new DoubleCell(err) };
}
@Override
public DataColumnSpec[] getColumnSpecs() {
DataColumnSpecCreator crea = new DataColumnSpecCreator("PolyReg prediction", DoubleCell.TYPE);
DataColumnSpec col1 = crea.createSpec();
crea = new DataColumnSpecCreator("Prediction Error", DoubleCell.TYPE);
DataColumnSpec col2 = crea.createSpec();
return new DataColumnSpec[] { col1, col2 };
}
@Override
public void setProgress(final int curRowNr, final int rowCount, final RowKey lastKey, final ExecutionMonitor execMon) {
// do nothing
}
};
}
use of org.knime.core.data.DataColumnSpecCreator in project knime-core by knime.
the class RegressionPredictorNodeModel method createRearranger.
private ColumnRearranger createRearranger(final DataTableSpec inSpec, final PMMLPortObjectSpec regModelSpec, final PMMLRegressionTranslator regModel) throws InvalidSettingsException {
if (regModelSpec == null) {
throw new InvalidSettingsException("No input");
}
// exclude last (response column)
String targetCol = "Response";
for (String s : regModelSpec.getTargetFields()) {
targetCol = s;
break;
}
final List<String> learnFields;
if (regModel != null) {
RegressionTable regTable = regModel.getRegressionTable();
learnFields = new ArrayList<String>();
for (NumericPredictor p : regTable.getVariables()) {
learnFields.add(p.getName());
}
} else {
learnFields = new ArrayList<String>(regModelSpec.getLearningFields());
}
final int[] colIndices = new int[learnFields.size()];
int k = 0;
for (String learnCol : learnFields) {
int index = inSpec.findColumnIndex(learnCol);
if (index < 0) {
throw new InvalidSettingsException("Missing column for " + "regressor variable : \"" + learnCol + "\"");
}
DataColumnSpec regressor = inSpec.getColumnSpec(index);
String name = regressor.getName();
DataColumnSpec col = inSpec.getColumnSpec(index);
if (!col.getType().isCompatible(DoubleValue.class)) {
throw new InvalidSettingsException("Incompatible type of " + "column \"" + name + "\": " + col.getType());
}
colIndices[k++] = index;
}
// try to use some smart naming scheme for the append column
String oldName = targetCol;
if (inSpec.containsName(oldName) && !oldName.toLowerCase().endsWith("(prediction)")) {
oldName = oldName + " (prediction)";
}
String newColName = DataTableSpec.getUniqueColumnName(inSpec, oldName);
DataColumnSpec newCol = new DataColumnSpecCreator(newColName, DoubleCell.TYPE).createSpec();
SingleCellFactory fac = new SingleCellFactory(newCol) {
@Override
public DataCell getCell(final DataRow row) {
RegressionTable t = regModel.getRegressionTable();
int j = 0;
double result = t.getIntercept();
for (NumericPredictor p : t.getVariables()) {
DataCell c = row.getCell(colIndices[j++]);
if (c.isMissing()) {
return DataType.getMissingCell();
}
double v = ((DoubleValue) c).getDoubleValue();
if (p.getExponent() != 1) {
v = Math.pow(v, p.getExponent());
}
result += p.getCoefficient() * v;
}
return new DoubleCell(result);
}
};
ColumnRearranger c = new ColumnRearranger(inSpec);
c.append(fac);
return c;
}
use of org.knime.core.data.DataColumnSpecCreator in project knime-core by knime.
the class MissingValueHandling3Table method createTableSpecPrivate.
/* private helper that assumes the ColSetting to have the right format. */
private static DataTableSpec createTableSpecPrivate(final DataTableSpec spec, final MissingValueHandling2ColSetting[] sets) {
assert (spec.getNumColumns() == sets.length);
DataColumnSpec[] newSpecs = new DataColumnSpec[sets.length];
for (int i = 0; i < sets.length; i++) {
DataColumnSpec colSpec = spec.getColumnSpec(i);
DataColumnSpec newSpec = colSpec;
if (sets[i].getMethod() == MissingValueHandling2ColSetting.METHOD_FIX_VAL) {
DataColumnDomain dom = colSpec.getDomain();
Comparator<DataCell> comp = colSpec.getType().getComparator();
DataCell fixCell = sets[i].getFixCell();
boolean changed = false;
DataCell l = dom.getLowerBound();
// (but rather be null). It may happen anyway, we catch it here
if (l != null && !l.isMissing() && (comp.compare(fixCell, l) < 0)) {
changed = true;
l = fixCell;
}
DataCell u = dom.getUpperBound();
if (u != null && !u.isMissing() && (comp.compare(fixCell, u) > 0)) {
changed = true;
u = fixCell;
}
Set<DataCell> vals = dom.getValues();
if (vals != null && !vals.contains(fixCell)) {
changed = true;
vals = new LinkedHashSet<DataCell>(vals);
vals.add(fixCell);
}
if (changed) {
DataColumnDomain newDom = new DataColumnDomainCreator(vals, l, u).createDomain();
DataColumnSpecCreator c = new DataColumnSpecCreator(colSpec);
c.setDomain(newDom);
newSpec = c.createSpec();
}
}
newSpecs[i] = newSpec;
}
return new DataTableSpec(newSpecs);
}
use of org.knime.core.data.DataColumnSpecCreator in project knime-core by knime.
the class Rule method main.
/**
* Zum Testen...
*
* @param args Pieps
* @throws Exception Tröt
*/
public static void main(final String[] args) throws Exception {
DataColumnSpec[] colSpecs = { new DataColumnSpecCreator("A", IntCell.TYPE).createSpec(), new DataColumnSpecCreator("B", IntCell.TYPE).createSpec(), new DataColumnSpecCreator("C", IntCell.TYPE).createSpec(), new DataColumnSpecCreator("S", StringCell.TYPE).createSpec(), new DataColumnSpecCreator("X", DoubleCell.TYPE).createSpec(), new DataColumnSpecCreator("Y", DoubleCell.TYPE).createSpec(), new DataColumnSpecCreator("Z", DoubleCell.TYPE).createSpec() };
DataTableSpec ts = new DataTableSpec(colSpecs);
BufferedReader in = new BufferedReader(new InputStreamReader(System.in));
String line;
while ((line = in.readLine()) != null) {
try {
Rule r = new Rule(line, ts);
System.out.println(r.toString());
} catch (ParseException ex) {
ex.printStackTrace();
}
}
}
use of org.knime.core.data.DataColumnSpecCreator in project knime-core by knime.
the class PMMLDataDictionaryTranslator method addColSpecsForDataFields.
/**
* @param pmmlDoc the PMML document to analyze
* @param colSpecs the list to add the data column specs to
*/
private void addColSpecsForDataFields(final PMMLDocument pmmlDoc, final List<DataColumnSpec> colSpecs) {
DataDictionary dict = pmmlDoc.getPMML().getDataDictionary();
for (DataField dataField : dict.getDataFieldArray()) {
String name = dataField.getName();
DataType dataType = getKNIMEDataType(dataField.getDataType());
DataColumnSpecCreator specCreator = new DataColumnSpecCreator(name, dataType);
DataColumnDomain domain = null;
if (dataType.isCompatible(NominalValue.class)) {
Value[] valueArray = dataField.getValueArray();
DataCell[] cells;
if (DataType.getType(StringCell.class).equals(dataType)) {
if (dataField.getIntervalArray().length > 0) {
throw new IllegalArgumentException("Intervals cannot be defined for Strings.");
}
cells = new StringCell[valueArray.length];
if (valueArray.length > 0) {
for (int j = 0; j < cells.length; j++) {
cells[j] = new StringCell(valueArray[j].getValue());
}
}
domain = new DataColumnDomainCreator(cells).createDomain();
}
} else if (dataType.isCompatible(DoubleValue.class)) {
Double leftMargin = null;
Double rightMargin = null;
Interval[] intervalArray = dataField.getIntervalArray();
if (intervalArray != null && intervalArray.length > 0) {
Interval interval = dataField.getIntervalArray(0);
leftMargin = interval.getLeftMargin();
rightMargin = interval.getRightMargin();
} else if (dataField.getValueArray() != null && dataField.getValueArray().length > 0) {
// try to derive the bounds from the values
Value[] valueArray = dataField.getValueArray();
List<Double> values = new ArrayList<Double>();
for (int j = 0; j < valueArray.length; j++) {
String value = "";
try {
value = valueArray[j].getValue();
values.add(Double.parseDouble(value));
} catch (Exception e) {
throw new IllegalArgumentException("Skipping domain calculation. " + "Value \"" + value + "\" cannot be cast to double.");
}
}
leftMargin = Collections.min(values);
rightMargin = Collections.max(values);
}
if (leftMargin != null && rightMargin != null) {
// set the bounds of the domain if available
DataCell lowerBound = null;
DataCell upperBound = null;
if (DataType.getType(IntCell.class).equals(dataType)) {
lowerBound = new IntCell(leftMargin.intValue());
upperBound = new IntCell(rightMargin.intValue());
} else if (DataType.getType(DoubleCell.class).equals(dataType)) {
lowerBound = new DoubleCell(leftMargin);
upperBound = new DoubleCell(rightMargin);
}
domain = new DataColumnDomainCreator(lowerBound, upperBound).createDomain();
} else {
domain = new DataColumnDomainCreator().createDomain();
}
}
specCreator.setDomain(domain);
colSpecs.add(specCreator.createSpec());
m_dictFields.add(name);
}
}
Aggregations