Examples of org.encog.ml.data.MLDataPair

org.encog.ml.data.MLDataPair
Training data is stored in two ways, depending on if the data is for supervised, or unsupervised training. For unsupervised training just an input value is provided, and the ideal output values are null. For supervised training both input and the expected ideal outputs are provided. This interface abstracts classes that provide a holder for both of these two data items. @author jheaton

    temporal.generate();
    
    Iterator<MLDataPair> itr = temporal.iterator();
    
    // set 0
    MLDataPair pair = itr.next();
    TestCase.assertEquals(10, pair.getInput().size());
    TestCase.assertEquals(1, pair.getIdeal().size());
    TestCase.assertEquals(3.0, pair.getInput().getData(0));
    TestCase.assertEquals(1.5, pair.getInput().getData(1));
    TestCase.assertEquals(3.0, pair.getInput().getData(2));
    TestCase.assertEquals(0.6, pair.getInput().getData(3));
    TestCase.assertEquals(3.0, pair.getInput().getData(4));
    TestCase.assertEquals(0.375, pair.getInput().getData(5));
    TestCase.assertEquals(3.0, pair.getInput().getData(6));
    TestCase.assertEquals(0.25, Math.round(pair.getInput().getData(7)*4.0)/4.0);
    TestCase.assertEquals(3.0, pair.getInput().getData(8));
    TestCase.assertEquals(0.25, Math.round(pair.getInput().getData(9)*4.0)/4.0);
    TestCase.assertEquals(18.0, pair.getIdeal().getData(0));
  }

View Full Code Here

      }
    }


    this.network.setExclude((int) training.getRecordCount());


    final MLDataPair pair = BasicMLDataPair.createPair(
        training.getInputSize(), training.getIdealSize());


    final double[] out = new double[this.network.getOutputCount()];


    for (int r = 0; r < training.getRecordCount(); r++) {
      training.getRecord(r, pair);
      this.network.setExclude(this.network.getExclude() - 1);


      err = 0.0;


      final MLData input = pair.getInput();
      final MLData target = pair.getIdeal();


      if (this.network.getOutputMode() == PNNOutputMode.Unsupervised) {
        if (deriv) {
          final MLData output = computeDeriv(input, target);
          for (int z = 0; z < this.network.getOutputCount(); z++) {
            out[z] = output.getData(z);
          }
        } else {
          final MLData output = this.network.compute(input);
          for (int z = 0; z < this.network.getOutputCount(); z++) {
            out[z] = output.getData(z);
          }
        }
        for (int i = 0; i < this.network.getOutputCount(); i++) {
          diff = input.getData(i) - out[i];
          err += diff * diff;
        }
      } else if (this.network.getOutputMode() == PNNOutputMode.Classification) {
        final int tclass = (int) target.getData(0);
        MLData output;


        if (deriv) {
          output = computeDeriv(input, pair.getIdeal());
        } else {
          output = this.network.compute(input);
        }


        EngineArray.arrayCopy(output.getData(),out);

View Full Code Here

        this.v[vsptr + ivar] = 0.0;
        this.w[wsptr + ivar] = 0.0;
      }
    }


    final MLDataPair pair = BasicMLDataPair.createPair(this.network
        .getSamples().getInputSize(), this.network.getSamples()
        .getIdealSize());


    for (int r = 0; r < this.network.getSamples().getRecordCount(); r++) {


      this.network.getSamples().getRecord(r, pair);


      if (r == this.network.getExclude()) {
        continue;
      }


      dist = 0.0;
      for (ivar = 0; ivar < this.network.getInputCount(); ivar++) {
        diff = input.getData(ivar) - pair.getInput().getData(ivar);
        diff /= this.network.getSigma()[ivar];
        this.dsqr[ivar] = diff * diff;
        dist += this.dsqr[ivar];
      }


      if (this.network.getKernel() == PNNKernelType.Gaussian) {
        dist = Math.exp(-dist);
      } else if (this.network.getKernel() == PNNKernelType.Reciprocal) {
        dist = 1.0 / (1.0 + dist);
      }


      truedist = dist;
      if (dist < 1.e-40) {
        dist = 1.e-40;
      }


      if (this.network.getOutputMode() == PNNOutputMode.Classification) {
        pop = (int) pair.getIdeal().getData(0);
        out[pop] += dist;
        vptr = pop * this.network.getInputCount();
        wptr = pop * this.network.getInputCount();
        for (ivar = 0; ivar < this.network.getInputCount(); ivar++) {
          temp = truedist * this.dsqr[ivar];
          this.v[vptr + ivar] += temp;
          this.w[wptr + ivar] += temp * (2.0 * this.dsqr[ivar] - 3.0);
        }
      }


      else if (this.network.getOutputMode() == PNNOutputMode.Unsupervised) {
        for (ivar = 0; ivar < this.network.getInputCount(); ivar++) {
          out[ivar] += dist * pair.getInput().getData(ivar);
          temp = truedist * this.dsqr[ivar];
          this.v[vsptr + ivar] += temp;
          this.w[wsptr + ivar] += temp
              * (2.0 * this.dsqr[ivar] - 3.0);
        }
        vptr = 0;
        wptr = 0;
        for (outvar = 0; outvar < this.network.getOutputCount(); outvar++) {
          for (ivar = 0; ivar < this.network.getInputCount(); ivar++) {
            temp = truedist * this.dsqr[ivar]
                * pair.getInput().getData(ivar);
            this.v[vptr++] += temp;
            this.w[wptr++] += temp * (2.0 * this.dsqr[ivar] - 3.0);
          }
        }
        psum += dist;
      } else if (this.network.getOutputMode() == PNNOutputMode.Regression) {


        for (ivar = 0; ivar < this.network.getOutputCount(); ivar++) {
          out[ivar] += dist * pair.getIdeal().getData(ivar);
        }
        vptr = 0;
        wptr = 0;
        for (outvar = 0; outvar < this.network.getOutputCount(); outvar++) {
          for (ivar = 0; ivar < this.network.getInputCount(); ivar++) {
            temp = truedist * this.dsqr[ivar]
                * pair.getIdeal().getData(outvar);
            this.v[vptr++] += temp;
            this.w[wptr++] += temp * (2.0 * this.dsqr[ivar] - 3.0);
          }
        }
        for (ivar = 0; ivar < this.network.getInputCount(); ivar++) {

View Full Code Here

    return data.size();
  }


  @Override
  public void getRecord(long index, MLDataPair pair) {
    final MLDataPair source = this.data.get((int) index);
    pair.setInputArray(source.getInputArray());
    if (pair.getIdealArray() != null) {
      pair.setIdealArray(source.getIdealArray());
    }
  }

View Full Code Here

      oseqIterator.next();
    }


    int t = 1;
    while (oseqIterator.hasNext()) {
      final MLDataPair observation = oseqIterator.next();


      for (int i = 0; i < hmm.getStateCount(); i++) {
        computeStep(hmm, observation, t, i);
      }

View Full Code Here


    this.currentState = this.hmm.getStateCount() - 1;
  }


  public MLDataPair observation() {
    final MLDataPair o = this.hmm.getStateDistribution(this.currentState)
        .generate();
    double rand = Math.random();


    for (int j = 0; j < (this.hmm.getStateCount() - 1); j++) {
      if ((rand -= this.hmm

View Full Code Here

   * {@inheritDoc}
   */
  @Override
  public MLDataPair next() {
    if (hasNext()) {
      final MLDataPair pair = BasicMLDataPair.createPair(
          this.owner.getInputSize(), this.owner.getIdealSize());
      this.owner.getRecord(this.currentIndex++, pair);
      return pair;
    } else {
      return null;

View Full Code Here

    if (seqIterator.hasNext()) {
      seqIterator.next();
    }


    for (int t = 1; t < oseq.size(); t++) {
      final MLDataPair observation = seqIterator.next();


      for (int i = 0; i < hmm.getStateCount(); i++) {
        computeAlphaStep(hmm, observation, t, i);
      }
    }

View Full Code Here

    return (int)getRecordCount();
  }


  @Override
  public MLDataPair get(int index) {
    MLDataPair result = BasicMLDataPair.createPair(getInputSize(), getIdealSize());
    this.getRecord(index, result);
    return result;
  }

View Full Code Here

    if (seqIterator.hasNext()) {
      seqIterator.next();
    }


    for (int t = 1; t < oseq.size(); t++) {
      final MLDataPair observation = seqIterator.next();


      for (int i = 0; i < hmm.getStateCount(); i++) {
        computeAlphaStep(hmm, observation, t, i);
      }
      scale(this.ctFactors, this.alpha, t);

View Full Code Here

0 1 2 3 4 5 6 7 8

TOP

Related Classes of org.encog.ml.data.MLDataPair

org.encog.app.analyst.util.AnalystUtility

org.encog.ensemble.adaboost.AdaBoost

org.encog.ensemble.data.EnsembleDataSet

org.encog.examples.clustering.kmeans.SimpleKMeans

org.encog.examples.neural.benchmark.BinaryVsMemory

org.encog.ml.data.auto.AutoFloatDataSet

org.encog.ml.data.auto.TestAutoFloatDataSet

org.encog.ml.data.basic.BasicMLDataPair

org.encog.ml.data.basic.BasicMLDataSet

org.encog.ml.data.basic.BasicMLSequenceSet

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.