Examples of Auc

edu.uci.jforestsx.eval.AUC
@author Yasser Ganjisaffar
org.apache.mahout.classifier.evaluation.Auc
Computes AUC and a few other accuracy statistics without storing huge amounts of data. This is done by keeping uniform samples of the positive and negative scores. Then, when AUC is to be computed, the remaining scores are sorted and a rank-sum statistic is used to compute the AUC. Since AUC is invariant with respect to down-sampling of either positives or negatives, this is close to correct and is exactly correct if maxBufferSize or fewer positive and negative scores are examined.

Examples of org.apache.mahout.classifier.evaluation.Auc

      if (!showAuc && !showConfusion && !showScores) {
        showAuc = true;
        showConfusion = true;
      }


      Auc collector = new Auc();
      LogisticModelParameters lmp = LogisticModelParameters.loadFrom(new File(modelFile));


      CsvRecordFactory csv = lmp.getCsvRecordFactory();
      OnlineLogisticRegression lr = lmp.createRegression();
      BufferedReader in = TrainLogistic.open(inputFile);
      String line = in.readLine();
      csv.firstLine(line);
      line = in.readLine();
      if (showScores) {
        output.printf(Locale.ENGLISH, "\"%s\",\"%s\",\"%s\"\n", "target", "model-output", "log-likelihood");
      }
      while (line != null) {
        Vector v = new SequentialAccessSparseVector(lmp.getNumFeatures());
        int target = csv.processLine(line, v);


        double score = lr.classifyScalar(v);
        if (showScores) {
          output.printf(Locale.ENGLISH, "%d,%.3f,%.6f\n", target, score, lr.logLikelihood(target, v));
        }
        collector.add(target, score);
        line = in.readLine();
      }


      if (showAuc) {
        output.printf(Locale.ENGLISH, "AUC = %.2f\n", collector.auc());
      }
      if (showConfusion) {
        Matrix m = collector.confusion();
        output.printf(Locale.ENGLISH, "confusion: [[%.1f, %.1f], [%.1f, %.1f]]\n",
          m.get(0, 0), m.get(1, 0), m.get(0, 1), m.get(1, 1));
        m = collector.entropy();
        output.printf(Locale.ENGLISH, "entropy: [[%.1f, %.1f], [%.1f, %.1f]]\n",
          m.get(0, 0), m.get(1, 0), m.get(0, 1), m.get(1, 1));
      }
    }
  }

View Full Code Here

Examples of org.apache.mahout.classifier.evaluation.Auc

      for (int pass = 0; pass < 20; pass++) {
        for (TelephoneCall observation : train) {
          lr.train(observation.getTarget(), observation.asVector());
        }
        if (pass % 5 == 0) {
          Auc eval = new Auc(0.5);
          for (TelephoneCall testCall : test) {
            eval.add(testCall.getTarget(), lr.classifyScalar(testCall.asVector()));
          }
          System.out.printf("%d, %.4f, %.4f\n", pass, lr.currentLearningRate(), eval.auc());
        }
      }


    }
  }

View Full Code Here

Examples of org.apache.mahout.classifier.evaluation.Auc

 */
public class NonWeightedAUCCrossValLossFunction implements CrossValLossFunction<PredictionMap> {


    @Override
    public double getLoss(List<LabelPredictionWeight<PredictionMap>> labelPredictionWeights)  {
        Auc auc = new Auc();
        for (LabelPredictionWeight<PredictionMap> labelPredictionWeight : labelPredictionWeights) {
            int trueValue = (Double) labelPredictionWeight.getLabel() == 1.0 ? 1 : 0;
            PredictionMap classifierPrediction = labelPredictionWeight.getPrediction();
            double probabilityOfCorrectInstance = classifierPrediction.get(labelPredictionWeight.getLabel());
            double score = 0;
            score = (trueValue == 1) ? probabilityOfCorrectInstance: 1 - probabilityOfCorrectInstance;
            auc.add(trueValue, score);
        }
        return 1 - auc.auc();
    }

View Full Code Here

Examples of org.apache.mahout.classifier.evaluation.Auc

        }
        crossValLoss.sortDataByProbability(aucDataList);
        ArrayList<WeightedAUCCrossValLossFunction.AUCPoint> aucPoints = crossValLoss.getAUCPointsFromData(aucDataList);
        double aucCrossValLoss = crossValLoss.getAUCLoss(aucPoints);


        Auc auc = new Auc();
        for(WeightedAUCCrossValLossFunction.AUCData aucData : aucDataList) {
            auc.add("test1".equals(aucData.getClassification()) ? 1 : 0, aucData.getProbability());
        }


        double mahoutAucLoss = 1.0 - auc.auc();
        //These aren't matching exactly, but the difference is minimal
        double acceptableDifference = 0.000000000001;
        Assert.assertTrue(Math.abs(mahoutAucLoss - aucCrossValLoss) < acceptableDifference);
    }

View Full Code Here

0 1

TOP

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.