Examples of cc.mallet.types.Instance

cc.mallet.types.Instance
A machine learning "example" to be used in training, testing or performance of various machine learning algorithms.
An instance contains four generic fields of predefined name: "data", "target", "name", and "source". "Data" holds the data represented `by the instance, "target" is often a label associated with the instance, "name" is a short identifying name for the instance (such as a filename), and "source" is human-readable sourceinformation, (such as the original text).
Each field has no predefined type, and may change type as the instance is processed. For example, the data field may start off being a string that represents a file name and then be processed by a {@link cc.mallet.pipe.Pipe} into a CharSequencerepresenting the contents of the file, and eventually to a feature vector holding indices into an {@link cc.mallet.types.Alphabet} holding words found in the file.It is up to each pipe which fields in the Instance it modifies; the most common case is that the pipe modifies the data field.
Generally speaking, there are two modes of operation for Instances. (1) An instance gets created and passed through a Pipe, and the resulting data/target/name/source fields are used. This is generally done for training instances. (2) An instance gets created with raw values in its slots, then different users of the instance call newPipedCopy() with their respective different pipes. This might be done for test instances at "performance" time.
Rather than store an {@link cc.mallet.types.Alphabet} in the Instance,we obtain it through the Pipe instance variable, because the Pipe also indicates where the data came from and how to interpret the Alphabet.
Instances can be made immutable if locked. Although unlocked Instances are mutable, typically the only code that changes the values in the four slots is inside Pipes.
Note that constructing an instance with a pipe argument means "Construct the instance and then run it through the pipe". {@link cc.mallet.types.InstanceList} uses this methodwhen adding instances through a pipeInputIterator. @see Pipe @see Alphabet @see InstanceList @author Andrew McCallum mccallum@cs.umass.edu

      double value = 0.0;
      Iterator<Instance> iter = trainingList.iterator();
      int ii=0;
      while (iter.hasNext()) {
        ii++;
        Instance instance = iter.next();
        double instanceWeight = trainingList.getInstanceWeight(instance);
        Labeling labeling = instance.getLabeling ();
        if (labeling == null)
          continue;
        //System.out.println("L Now "+inputAlphabet.size()+" regular features.");


        this.theClassifier.getClassificationScores (instance, scores);
        FeatureVector fv = (FeatureVector) instance.getData ();
        int li = labeling.getBestIndex();
        value = - (instanceWeight * Math.log (scores[li]));
        if(Double.isNaN(value)) {
          logger.fine ("MaxEntTrainer: Instance " + instance.getName() +
              "has NaN value. log(scores)= " + Math.log(scores[li]) +
              " scores = " + scores[li] + 
              " has instance weight = " + instanceWeight);


        }
        if (Double.isInfinite(value)) {
          logger.warning ("Instance "+instance.getSource() + " has infinite value; skipping value and gradient");
          cachedValue -= value;
          cachedValueStale = false;
          return -value;
//          continue;
        }

View Full Code Here

  /** Pipe the object through this classifier's pipe, then classify the resulting instance. */
  public Classification classify (Object obj)
  {
    if (obj instanceof Instance)
      return classify ((Instance)obj);
    return classify (instancePipe.instanceFrom(new Instance (obj, null, null, null)));
  }

View Full Code Here

  public double getAverageRank ()
  {
    double rsum = 0;
    Labeling tmpL;
    Classification tmpC;
    Instance tmpI;
    Label tmpLbl, tmpLbl2;
    int tmpInt;
    for(int i = 0; i < this.size(); i++) {
      tmpC = this.get(i);
      tmpI = tmpC.getInstance();
      tmpL = tmpC.getLabeling();
      tmpLbl = (Label)tmpI.getTarget();
      tmpInt = tmpL.getRank(tmpLbl);
      tmpLbl2 = tmpL.getLabelAtRank(0);
      rsum = rsum + tmpInt;
    }
    return rsum/this.size();

View Full Code Here

      // Initialize the constraints, using only the constraints from
      // the "positive" instance
      Iterator<Instance> iter = trainingList.iterator ();
      logger.fine("Number of instances in training list = " + trainingList.size());
      while (iter.hasNext()) {
        Instance instance = iter.next();
        double instanceWeight = trainingList.getInstanceWeight(instance);
        FeatureVectorSequence fvs = (FeatureVectorSequence) instance.getData();
        // label of best instance in subList
        Object target = instance.getTarget();
        Label label = null;
        if (target instanceof Labels)
          label = ((Labels)target).get(0);
        else label = (Label)target;
        int positiveIndex =
          Integer.valueOf(label.getBestLabel().getEntry().toString()).intValue();
        if (positiveIndex == -1) { // invalid instance
          logger.warning("True label is -1. Skipping...");
           continue;
        }
        FeatureVector fv = (FeatureVector)fvs.get(positiveIndex);
        Alphabet fdict = fv.getAlphabet();
        assert (fv.getAlphabet() == fd);


        // xxx ensure dimensionality of constraints correct
        MatrixOps.rowPlusEquals (constraints, numFeatures, 0, fv, instanceWeight);


        // For the default feature, whose weight is 1.0
        assert(!Double.isNaN(instanceWeight)) : "instanceWeight is NaN";
        //assert(!Double.isNaN(li)) : "bestIndex is NaN";
        boolean hasNaN = false;
        for(int i = 0; i < fv.numLocations(); i++) {
          if(Double.isNaN(fv.valueAtLocation(i))) {
            logger.info("NaN for feature " + fdict.lookupObject(fv.indexAtLocation(i)).toString()); 
            hasNaN = true;
          }
        }
        if(hasNaN)
          logger.info("NaN in instance: " + instance.getName());


        // default constraints for positive instances xxx
        constraints[0*numFeatures + defaultFeatureIndex] += 1.0 * instanceWeight;
      }
      //TestMaximizable.testValueAndGradientCurrentParameters (this);

View Full Code Here


    Iterator<Integer> keyIter = labeledFeatures.keySet().iterator();
    
    double[][] featureCounts = new double[labeledFeatures.size()][numLabels];
    for (int ii = 0; ii < trainingData.size(); ii++) {
      Instance instance = trainingData.get(ii);
      FeatureVector fv = (FeatureVector)instance.getData();
      Labeling labeling = trainingData.get(ii).getLabeling();
      double[] labelDist = new double[numLabels];
      
      if (labeling == null) {
        labelByVoting(labeledFeatures,instance,labelDist);

View Full Code Here

    int numLabels = list.getTargetAlphabet().size();
    
    double[][] featureLabelCounts = new double[numFeatures][numLabels];
    
    for (int ii = 0; ii < list.size(); ii++) {
      Instance instance = list.get(ii);
      FeatureVector featureVector = (FeatureVector)instance.getData();
      
      // this handles distributions over labels
      for (int li = 0; li < numLabels; li++) {
        double py = instance.getLabeling().value(li);
        for (int loc = 0; loc < featureVector.numLocations(); loc++) {
          int fi = featureVector.indexAtLocation(loc);
          double val;
          if (useValues) {
            val = featureVector.valueAtLocation(loc);

View Full Code Here

      if (m_ilist == null)
        throw new IllegalStateException ("Frozen.  Cannot split.");
      int numLeftChildren = 0;
      boolean[] toLeftChild = new boolean[m_instIndices.length];
      for (int i = 0; i < m_instIndices.length; i++) {
        Instance instance = m_ilist.get(m_instIndices[i]);
        FeatureVector fv = (FeatureVector) instance.getData();
        if (fv.value (m_gainRatio.getMaxValuedIndex()) <= m_gainRatio.getMaxValuedThreshold()) {
          toLeftChild[i] = true;
          numLeftChildren++;
        }
        else

View Full Code Here

      double value = 0.0;
      Iterator<Instance> iter = trainingList.iterator();
      int ii=0;
      while (iter.hasNext()) {
        ii++;
        Instance instance = iter.next();
        double instanceWeight = trainingList.getInstanceWeight(instance);
        Labeling labeling = instance.getLabeling ();
        if (labeling == null)
          continue;
        //System.out.println("L Now "+inputAlphabet.size()+" regular features.");


        this.theClassifier.getClassificationScores (instance, scores);
        FeatureVector fv = (FeatureVector) instance.getData ();


        value = 0.0;
        for(int pos = 0; pos < labeling.numLocations(); pos++) { //loop, added by Limin Yao
          int ll = labeling.indexAtLocation(pos);
          if (scores[ll] == 0  && labeling.valueAtLocation(pos) > 0) {
            logger.warning ("Instance "+instance.getSource() + " has infinite value; skipping value and gradient");
            cachedValue = Double.NEGATIVE_INFINITY;
            cachedValueStale = false;
            return cachedValue;
          }
          else if (labeling.valueAtLocation(pos) != 0) {
            value -= (instanceWeight * labeling.valueAtLocation(pos) * Math.log (scores[ll]));
          }
        }      


        if (Double.isNaN(value)) {
          logger.fine ("MaxEntOptimizableByLabelDistribution: Instance " + instance.getName() +
                 "has NaN value.");
        }
        if (Double.isInfinite(value)) {
          logger.warning ("Instance "+instance.getSource() + " has infinite value; skipping value and gradient");
          cachedValue -= value;
          cachedValueStale = false;
          return -value;
//          continue;
        }

View Full Code Here

    NeighborIterator iter = new ClusterSampleIterator(clustering,
                                                      random,
                                                      0.5,
                                                      10);
    while (iter.hasNext()) {
      Instance instance = (Instance)iter.next();
      System.err.println(instance.getData() + "\n");
    }
    
    System.err.println("\n\nPairSampleIterator");
    iter = new PairSampleIterator(clustering,
                                  random,
                                  0.5,
                                  10);
    while (iter.hasNext()) {
      Instance instance = (Instance)iter.next();
      System.err.println(instance.getData() + "\n");
    }


    System.err.println("\n\nAllPairsIterator");
    iter = new AllPairsIterator(clustering);                                
    while (iter.hasNext()) {
      Instance instance = (Instance)iter.next();
      System.err.println(instance.getData() + "\n");
    }
}

View Full Code Here

    double value = 0.;
    //double sumLogP = 0;
    
    for (int ii = 0; ii < trainingData.size(); ii++) {
      double[] scores = new double[numLabels];
      Instance instance = trainingData.get(ii);
      FeatureVector input = (FeatureVector) instance.getData ();
      double instanceWeight = trainingData.getInstanceWeight(ii);
      
      classifier.getClassificationScores(instance, scores);
      double logZ = Double.NEGATIVE_INFINITY;
      
      for (int li = 0; li < numLabels; li++) {
        if (baseDist != null && baseDist[ii][li] == 0) {
          scores[li] = Double.NEGATIVE_INFINITY;
        }
        else if (baseDist != null) {
          double logP = Math.log(baseDist[ii][li]);
          scores[li] += logP;  
        }
        logZ = Maths.sumLogProb(logZ, scores[li]);
      }
      assert(!Double.isNaN(logZ));
      value -= instanceWeight * logZ;
      
      if (Double.isNaN(value)) {
        logger.warning("Instance " + instance.getName() + " has NaN value.");
        continue;
      }
      if (Double.isInfinite(value)) {
        logger.warning("Instance " + instance.getName() + " has infinite value; skipping value and gradient");
        continue;
      }
      
      // exp normalize
      MatrixOps.expNormalize(scores);

View Full Code Here

0 1 2 3 4 5 6 7 8 9

TOP

Related Classes of cc.mallet.types.Instance

cc.mallet.classify.BalancedWinnowTrainer

cc.mallet.classify.C45$Node

cc.mallet.classify.Classification

cc.mallet.classify.Classifier

cc.mallet.classify.DecisionTree$Node

cc.mallet.classify.evaluate.ConfusionMatrix

cc.mallet.classify.FeatureConstraintUtil

cc.mallet.classify.MaxEntOptimizableByGE

cc.mallet.classify.MaxEntOptimizableByLabelDistribution

cc.mallet.classify.MaxEntOptimizableByLabelLikelihood

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.