Example usage for weka.attributeSelection ASEvaluation buildEvaluator

Introduction

In this page you can find the example usage for weka.attributeSelection ASEvaluation buildEvaluator.

Prototype

public abstract void buildEvaluator(Instances data) throws Exception;

Source Link

Document

Generates a attribute evaluator.

Usage

From source file:com.rapidminer.operator.features.weighting.GenericWekaAttributeWeighting.java

License:Open Source License

public AttributeWeights calculateWeights(ExampleSet exampleSet) throws OperatorException {
    AttributeWeights weights = new AttributeWeights();

    ASEvaluation evaluator = getWekaAttributeEvaluator(getOperatorClassName(),
            WekaTools.getWekaParametersFromTypes(this, wekaParameters));

    log("Converting to Weka instances.");
    Instances instances = WekaTools.toWekaInstances(exampleSet, "WeightingInstances",
            WekaInstancesAdaptor.WEIGHTING);
    try {/*from  w w  w  . j av  a2 s.  c o  m*/
        log("Building Weka attribute evaluator.");
        evaluator.buildEvaluator(instances);
        //evaluator.buildEvaluator(instances);
    } catch (UnassignedClassException e) {
        throw new UserError(this, e, 105, new Object[] { getOperatorClassName(), e });
    } catch (ArrayIndexOutOfBoundsException e) {
        throw new UserError(this, e, 105, new Object[] { getOperatorClassName(), e });
    } catch (Exception e) {
        throw new UserError(this, e, 905, new Object[] { getOperatorClassName(), e });
    }

    int index = 0;
    if (evaluator instanceof AttributeEvaluator) {
        AttributeEvaluator singleEvaluator = (AttributeEvaluator) evaluator;
        for (Attribute attribute : exampleSet.getAttributes()) {
            try {
                double result = singleEvaluator.evaluateAttribute(index++);
                weights.setWeight(attribute.getName(), result);
            } catch (Exception e) {
                logWarning("Cannot evaluate attribute '" + attribute.getName() + "', use unknown weight.");
            }
        }
    } else {
        logWarning("Cannot evaluate attributes, use unknown weights.");
    }

    return weights;
}

---

From source file:mulan.dimensionalityReduction.BinaryRelevanceAttributeEvaluator.java

License:Open Source License

/**
 * @param ase /*from  w w w  .ja v a 2 s  .co  m*/
 * @param mlData 
 * @param combapp combination approach mode ("max", "avg", "min")
 * @param norm normalization mode ("dl", "dm", "none")
 * @param mode scoring mode ("eval", "rank")
 */
public BinaryRelevanceAttributeEvaluator(ASEvaluation ase, MultiLabelInstances mlData, String combapp,
        String norm, String mode) {
    CombApprMode = combapp;
    NormMode = norm;
    ScoreMode = mode;

    numLabels = mlData.getNumLabels();
    int[] labelIndices = mlData.getLabelIndices();
    try {
        int numAttributes = mlData.getFeatureIndices().length;
        double[][] evaluations = new double[numLabels][numAttributes];
        for (int i = 0; i < numLabels; i++) {
            // create dataset
            Instances labelInstances = BinaryRelevanceTransformation.transformInstances(mlData.getDataSet(),
                    labelIndices, labelIndices[i]);

            // build evaluator
            ase.buildEvaluator(labelInstances);

            // evaluate features
            for (int j = 0; j < numAttributes; j++) {
                evaluations[i][j] = ((AttributeEvaluator) ase).evaluateAttribute(j);
            }
        }

        //scoring of features
        scores = featureSelection(evaluations);

    } catch (Exception ex) {
        ex.printStackTrace();
    }
}

---